PROCESS DESIGN COURSE VOLUME I PRESENTATION ZS EXXON ENGINEERINGPROCESS DESIGN COURSE VOLUMEI PRESENTATION EXXON ENGINEERINGis 1. 2. 3. DATA St Given: Given: Given: 05309101. Pom TRAINING PERFORMANCE OBJECTIVES EXXON ENGINEERING PROCESS DESIGN COURSE SOURCES + A.gas stream of three defined components + The mol. fraction of each component + Temperature and pressure of gas stream Determine the following properties of the stream: Molecular weight Weight and mola] average boiling points Pseudocritical temperature Pseudocritical pressure Absolute viscosity Compressibility factor Density in pounds/ft3 at conditions. eteaoce + A liquid stream of three defined components * Volume fraction of each component + Temperature and pressure of liquid stream Determine the following properties of the stream: Molecular weight Mean, molal, volume, and weight average boiling points Characterization factor True critical temperature True critical pressure Density in pounds/ft3 at 60°F Density in *API at 60°F eroance * A liquid stream of three defined components * The weight fraction of each component. * Viscosity of each component at 100°F and 210°F Determine: a - The viscosity of the mixture at 100°F and 210°F via the Blending Index procedure. b - The viscosity of the mixture at 150°F using viscosity-temperature charts. c- The difference in enthalpy between the stream as a liquid at 100°F and 30 atmospheres and the stream as a vapor at 600°F and 1 atmosphere. d- The latent heat of vaporization at 600°F.4. 5. Given 6. Given: . Given: 05309101.POM + A gas stream of three defined components + The mol fraction of each component + Temperature and pressure of the gas stream Flow rate of stream in MegaSCF/SD Determine: a b - The brake horsepower required to compress the stream adiabatically to twice the given pressure - Temperature of the compressed stream +A stream of three defined components + The mol fraction of each component + Pressure of the stream Estimate the following properties of the stream using Vapor Pressures: a b * A stream of three defined components - Bubble point temperature - Dew point temperature The mol fraction of each component The equilibrium instants for each component + Temperature and pressure of the stream Determine the following properties of the stream: a b - Mole percent liquid at given conditions = Liquid and vapor phase compositions ASTM Distillation of pipestil] gas oil cut API gravity Determine: a - Specific heat of stream at 200°F, 300°F, and 400°F (correct for Ky) - Latent heat of vaporization of stream at 200°F, 300°F, and 400°F - 15/5 Distillation Curve - The "names" of an arbitrary number of pseudocomponents that will define the stream CxDISTILLAT! (OLOGY AND TOWER DESIGN 1. Given: + Temperature, pressure, and rate of C3-430°F refinery stream + Composition of the stream through C5 material + Product specifications for Cq in the bottoms and Cs in the overhead * ASTM distillation of the C5+ material Conditions of utilities that are available Be able to - Determine a quickie preliminary design using Fenske-Gilliland procedure 2. Be able to define the following tray design terms: a - Blowing g - Single Pass Tray m - Downcomer Clearance b - Dumping h - Bubble Area n - Two Pass Tray ¢ - Jet Flooding i - Hole Area o - Antijump Baffle d - Liquid Gradient dj - Perforated Area P - Downcomer Seal e - Weeping k - Waste Area @ - Downcomer Filling f - Ultimate Capacity 1 Free Area 3. Be able to classify the following types of trays: - Bubble Cap - det ~ Sieve - Valve In order of increasing: a - Capacity c - Fractionation Efficiency b - Cost per Unit Area d - Flexibility 4. Given: ~ A liquid stream’ s-- + Flow rate (pounds/hour) and composition + Surface tension * Viscosity + Temperature - A vapor stream’s -~ * Flow rate (pounds/hour) and composition + Viscosity + Temperature + Pressure - Required flexibility of the service Be able to design a sieve tray by hand for this serviceIIL. HEAT TRANSFER FUNDAMENTALS 1. Given: + Shell and tube exchangers for concurrent, countercurrent, and mixed flow Be able to: a - Identify each b - Explain the differences among them 2. Given: + Temperature in and out for both sides of S & T exchanger + Number of tube and shell side passes «No change in phase for either fluids Be able to: a - Calculate the ATLM b - Calculate Fn ¢ - State the temperature approach d - State the temperature cross e - Explain why one should not design an exchanger that results in outlet temperatures giving Fy <0.8 3. Given: * Temperatures in and out for both sides of a water condenser * 1/Q curve for the HC being condensed Be able to: - Calculate the effective temperature driving force (Ate) 4. Given: « Temperatures in and out for both sides of S & T water cooler * Rate of HC on the shell side + API gravity and MABP of the HC + Exchanger geometry (tube sizes, area, etc.) Be able to: Calculate the individual film coefficients and the overall coefficient - Calculate the total resistance due to fouling - Calculate tube wall temperature on the water side Specify which, if either, film coefficient is controlling - Estimate the effect on the overall coefficient if the rate of HC is doubled eane 05309101.POM5. \T_EXCHANGER DESIGN CONSIDERATIONS Given: Be ab] Be abl and wa‘ Given: Given: (05309101. POM * An exchanger design with less than the maximum allowable AP on both sides Be able to: a - Explain why low pressure drop is undesirable b - List steps that could be taken to increase AP fe to list two advantages for double-pipe exchangers. e to list the factors that enter into the choice between air-cooled ter-cooled exchangers. + The total duty of a $ & T exchanger (Btu/h) + The characteristics at the two fluids in the exchanger « The temperatures and pressures of both fluids + Type of service * Tube side and shel] side fouling factors Be able to: a - Determine the minimum number of shells and explain reason for this minimum b - Determine which fluid should be on the tube side c - Specify whether a fixed tube sheet, U-tube, or pull-through floating head design should be used d - Specify the tube layout - Specify tube type (bare or low fin) © The task of designing aS & T exchanger with specified shell side pressure drop Be able to: Specify minimum and maximun baffle spacing Explain reason for maximum limitation ¢ - Specify the most desirable baffle orientation if the exchanger has condensing vapor on shell side List advantages and disadvantages for high pressure drop through the shell side a6. Given: Shell side hydrocarbon rate (no change of phase) Shel] side fluid properties Temperatures and pressures Allowable pressure drop Water side inlet temperature Water side allowable AP Be able to: - Estimate the area of the exchanger 5309101 .POMe Vv. FURNACES 1. Given 2. Given + The design of a furnace from a furnace supplier + The process information and requirements for the furnace + Type of fuel and percent excess air Be able to: a - Determine radiant heat flux b - Calculate heat transferred in the radiant section ¢ - Calculate bridgewall temperature d - Calculate heat transferred in the convection section e - Calculate the stack temperature f - Calculate maximum tube metal temperature in radiant section g - Calculate fin-tip temperature in the convection section h - Calculate film temperatures ‘* Fuel gas type and rate to a furnace « Stack temperature + Flue gas analysis + Anbient air temperature Be able to: a - Calculate the percent excess air b - Calculate the furnace efficiency 3. The following information is given for a 100 M Btu/hour (heat to oi1) furnace: 05309101. POM * 75% of heat load is obtained with 15° API fuel oi] + 25% of heat load is obtained with 1000 Btu/scf fuel gas O11 burners operate at 25% excess air * Gas burners operate at 15% excess air * 0.6 pounds of atomizing stean (150 psig saturated) is required per pound of fuel oi] fired * Stack temperature is 60°F Determine: a - Furnace efficiency b - Pounds per hour of flue gasVL. _FLUID FLOW, PUMPS, COMPRESSORS 1. Given + A piping system handling a liquid hydrocarbon stream + Stream flowrate and properties + Length and size of pipe «+ Number and type of fittings Be able to: a - Estimate equivalent length of pipe b - Estimate pressure drop 2. Be able to define the term "cavitation" 3. Given the following data for each of four pumping services: Rate Specific gravity and viscosity NPSH available to the pumps Solids content Differential head required Viscosity Be able to select from the following list the type of pump to be used in each service a - Centrifugal b - Rotary screw ¢ - Reciprocating 4. Given the following data for a pumping service: Horizontal centrifugal pump with motor driver Suction drum is an overhead distillate drum Elevation of drum and pump centerline Size, length, and number of fittings in the suction line Rate, specific gravity and viscosity of the liquid being pumped Be able to determine the NPSH available to the pump 05309101 .POM@ 5. Given the same data as TPO Number 4 plus * Pump characteristic, efficiency, and NPSH curves + Pump impeller diameter «© Flow control valve pressure drop « Piping system downstream of pump « Downstream controlled pressure Be able to determine fo b- c- Brake horsepower required by pump System resistance curve Diameter to which impeller should be trimmed to allow for a control valve pressure drop equal to 20% of the line friction pressure drop Be able to list process information required to specify pump and spare 6. Given, for a centrifugal compressor: « Composition and rate of gas stream feeding the compressor + Discharge pressure » Suction pressure and temperature Be able to estimate: ance 05309101.PON Temperature rise exponent Discharge temperature Polytropic compression exponent Brake horsepower-10- WIL. PROCESS CONTROL 1. For Item under Fluid Flow Section, be able to determine the control valve size. 2. Given 3. Given 4. Given 5. Be abl 05309101. Pom * A list of services such as feed from tankage to tower, fractionator reflux, etc. Be able to: a - List the four fundamental types of control b - Select the type or types of control which would be most frequently used for each service ¢ - Select the control valve action on instrument air failure for each contro] valve ‘An incomplete process design flowplow for a multipass process furnace Properties and operating conditions of process stream Fuel gas pressure Pilot gas source is natural gas (pipeline) Coil outlet temperature control is to be used Furnace pressure drop Be able to prepare sketch showing typical instrumentation for controlling furnace operation * Compressor suction knockout drum with automatic liquid drawoff to blowdown drum Be able to: ~ Sketch drum and compressor showing all instrumentation needed for the drum, including alarms and compressor shutdown je to: a - Define the term rangeability b - State the maximum rangeability for a control valve (double seated, globe) and for an orifice meter. © - Specify when a rotameter, turbine meter and venturi are used’ e -u- DRUM. Given: 05309101.POM * Existing P&ID for a tower that is to be expanded * Detailed Heat and Material Balance for the new conditions Be able to: a - Calculate the maximum velocity permitted in the vapor-liquid distillate separator drum - Determine the volume required in the drum for the liquid - Determine the length and diameter of the drum - Prepare a detailed process design drawing of the drum suitable for submission to a vendor aoc-12- IX. GENERAL PROCESS DESIGN CONSIDERATIONS 1. Be able to define the following terms: a - Operating temperature b - Maximum fluid temperature c - Design temperature d - Operating pressure e - Design pressure f - Short time design basis 2. Be able to explain why it is very important to minimize the difference betwee operat: 3. Given 4. Given Given 05309102. Pom n operating temperature and design temperature for carbon steel ‘ing above 650°F * The task of designing a depropanizer, a butane splitter and a gasoline rerun tower Be able to a - List the factors that should be considered in specifying the design pressures and the minimum design temperatures for the vessels. b - Explain how the vendor will know what internal pressure the bottom of the towers will have to withstand + Temperatures and pressures in and out of two exchangers in series in discharge from pump + Pump curves + Normal and maximum operating conditions upstream of the pump * Exchangers have bypasses Be able to a - Specify the design temperatures for each exchanger b - Specify the design pressure for each exchanger + A carbon steel piping system with uninsulated flanges * Design temperature + Design pressure Be able to a - Determine the flange rating for the pipe b - Calculate the maximum design pressure for the system using the selected flanges c - Calculate the maximum pressure to which the system may be subjected at the given temperature-B- e 6. In addition to the onsite process design considerations, be able to list at least three other systems that need to be considered in the complete design of a project. In the preparation of a plot plan layout, be able to specify the desirable location for: a - Furnaces with respect to other equipment b - Cooling towers with respect to the unit c - Air fin exchangers with respect to fractionating towers and furnaces d - Pumps with respect to fractionating towers and furnaces 05309101 POM-u- X._ SAFETY. CONSIDERATIONS 1, Be able to list the references used in the design of emergency release systems. 2. Be able to Sorusse-heance 3. Given © The process flow sheet for a unit Explain the meaning of the following Autoignition Temperature Contingency Emergency Explosive Limits Flash Point - High Flash Stocks - Light Ends Low Flash Stocks Pyrophoric Material Single Risk Toxic Material Fire Zone Pressure Relief Valve Spacing standards Be able to N&xE MeABP, °F (of Fraction) ——»CRITICAL PROPERTIES + Our Primary Use — Compressibility of Gases — Correlations for Equilibrium Constants + True-Critical - for Pure Compounds. + Pseudo-Critical - for Mixtures PSEUDO-CRITICAL TEMPERATURE (T,,) Light Hydrocarbons Only 400 ¢ 350 Ls Mixtures 3 Base Curve - = Mol ABP Vs Tpc & 300 Gr. Curve - WABP Vs Tyo Maxwell p. 70 Blue Book p. 6.16 0 25 50 75 100 Average Bolling Point °F ——»PSEUDO-CRITICAL TEMPERATURE (Tec) Petroleum Fractions APL 10 ar Tec V8 Mol. ABP Maxwoll p. 72 Blue Book p. 6.30 Critical Temp. Average Bolling Point For Narrow Bolling Fret. Vol. ABP = Mol. ABP PSEUDO-CRITICAL PRESSURE (Ppc) Light Hydrocarbons - Normal Paraffin Mixtures Poo Maxwoll p. 71 Blue Book p. 6.19 Molecular WeightPSEUDO-CRITICAL PRESSURE (Ppc) Petroleum Fractions ‘API Pre SK fe ese na MABP, °F GAS DENSITY PV =npRT = uRT for 1 Mol Volume of 1 Mol = HRT ‘Weight of 1 Mol = M =the M Pe ie aRT P=PSIA P T=R p= PM = Ibs HRT” te? 112DENSITY OF LIQUIDS + See Section 8 of Maxwell + Density Calculations Do Not Blend API. Blend Specific Gravity, or Ibs/t? THERMAL EXPANSION OF LIQUIDS 1000 Mol ABP Vol @T Vol. @ 60° ‘Maxwell Section 8 Blue Book Section 8 1.0, °FTHERMAL PROPERTIES ‘Section 7 of Maxwell and Blue Book + Specific Heat + Latent Heat + Enthalpy 145 VISCOSITY ‘Section 9 of Maxwell and Blue Book + Pure Compounds + Pet. Fractions + Visc. - Temp. Charts + Vise. Conversions + Vise. Blending + VBIVISCOSITY OF LIQUID MIXTURES Page 173 Maxwell Page 9.61 Blue Book Vise. @ Comp, Vol. Fr, 50°F. CS vel A 0s 15.0 36.4 B 03 31.5 31.8 c 22 131.0 247 1.0 IVF x VBI = 32.68 Vol. Average Viscosity,CS 26.7 447 VISCOSITY OF GAS MIXTURES + Viscosity Charts - Section 9 of Maxwell & Blue Book + Calculations: Za NeZs¥Mi + NaZ2Vil +. NVM + NoMa +... Z = Visc. of Mix. at Atmos. Press. Ny, No = Mol. Fr. of indiv. Components 21s Zz = Vise. of Indiv. Components M,, Mz = Mol. Wt of Components * Correct for Pressure (Maxwell Page 177, Blue Book 9.87) 118VAPOR-LIQUID EQUILIBRIUM + Whenever a Liquid and Vapor Exist in Equilibrium, It Is. Possible to Determine the Composition of Each of the Phases. V-L-E Relationships Are the Basis for Bubble Point, Dew Point and Flash Calculations. + For Any Specific Component at V-L Equilibrium, its Concentration in the Vapor Phase Divided by its Concentration in the Liquid Phase Is Known as its Equilibrium Constant, or K. IDEAL SYSTEM — EQUILIBRIUM CONSTANTS Raoult's Law: p, = P,X, Dalton's Law: 9, =nY, PX, =2Y, ‘NeoPM eK, xX K, = 4(T,P) = Partial Pressure of i in Vapor P, = Vapor Pressure of i @ Temp. ‘otal System Pressure Mol. Fret. of i in Liquid Y, = Mol. Fret. of iin Vapor K = Equilibrium Constant Good up to About 4 Atmospheres for Systems Comprised of Similar CompoundsNON-IDEAL SYSTEM + Activity Coetticlent Corrects for Non-Ideal Liquid Solutions — Neccessary In Calculations involving Polar Compounds; @.g. Sour Water Strippers + Equilibrium Constants for Non-Ideal Vapor Phase Are Obtained from a Varlety of Computerized Correlations, e.g. — Benedict Webb Rubin — Chao Seader — Redlich Kwong FUGACITY FUNCTIONS + Fugacity Function May be Considered a Corrected Vapor Pressure. + Honee K, = (Fd . Fugacity Function Total Pressure + When K, is Determined by a Rigorous Methods such as Chao - ‘Seader, then the product of K, and x may be Called a “Fugacity Function". + Maxwell Chapter 5 Gives Graphs for Pure Component Fugacity Functions. These Are Only Approximate Values (Non Composition- Dependent). They Cannot be Used for Definitive Designs.BUBBLE POINT CALCULATION USING GAS LAWS ‘System Pressure = 120 PSIA = 8.17 ATM =x (See Maxwell Chapter 4) Vapor Pressure = P, K= Pin 10E 18 14 BOE PE ur Cs 18.30 19.50 20.70 2.240 2300 «(2530 ey 7.46 8.00 851 0.913 0.979 1.040 ny 5.46 5.90 6.36 0.668 0.722 0.779 Y= Kix S_ 1e er Cs 0.0133 0.0298 0.0318 0.0336 Wy 0.9516 0.8688 0.9316 0.9897 nC, 9.0351 9.0234 9.0283 9.0273 1.0000, 0.9220 0.9887, 1.0506 Linear interpolation Gives Bubble Point = 126°F FUGB = 135.0°F CHAO = 137.5°F RKAZ = 135.8°F } maorou aod DEW POINT CALCULATION USING IDEAL GAS LAWS ‘System Pressure = 132 PSIA = 8.982 ATM = x Vapor Pressure = P, (SPE SSF 160 Cy 9.71 10.40 11.00 nCe 7.270 «(7.75 8.30 ICs 310 3.34 3.60 a IC, 0.4590 nC, 0.8393, ICs 9.0017 1.0000 K=Pi/n ASYE 185°F 1.061 1.160 0.809 «(0.863 0.345 = 0.372 X= WK; IPE 18 EOF 0.425 0.396 0.376 0.667 0.625 0.584 9.005 9.005 9.004 1.007 1.028 0.984 Linear interpolation Gives Bubble Point = 157°F FUGB = 157.5°F CHAOs 159.0°F } Rigorous Methods KAZ = 157.2°F 16°F 1.220 0.924EQUATIONS FOR FLASH VAPORIZATION F = Mols of Liquid Feed Vs Mols of Vapor Product L = Mols of Liquid Product 2, = Mol Fraction of i in Feed y= Mol Fraction Vapor X= Mol Fraction of i in Liquid + Material Balance: Fz, = Vy, + La %; +For 1 Mol of Feed: z, = Vy, + Lx, + By Equilibrium: =x = h 7 EQUATIONS FOR FLASH VAPORIZATION (Continued) 2, = Vy. + Ly (Previous Silde) Divide by vy,: Zt ate ety Le wi wy Ke ue wi(t+ ae) Vie 2 te VKFLASH VAPORIZATION CALCULATION tan lang ve > vie —_a_ a1 it+t VK STEP 1. Assume a Value of V. Then b= 1-V, and = 1:¥. STEP 2. Calculate Vy, for Each Component STEP 3, Sum to get V. iterate until Vessumed = Vestcutated Convergence Can be Difficult. Agreement of Vessumed With Vesicutated Should be within 0.1% to Assure a True Solution. FLASH CALCULATION EXAMPLE P=200PSIA T= 180°F a compe kn) 1+LvkK i+LWK c2 0.08 4.1710 1.3311 0.0601 c3 0.22 1.8650 1.7405 0.1264 NC4 0.53 0.8233 2.6774 0.1980 NCS 017 0.3760 4.6729 9.0364 1.00 0.4209 0.0964 0.4289 OKI (1) K Values Are From Rigorous Caloulation, Method RKJ2. 1.28DISTILLATION CONVERSIONS Definitions $155 — 15 Trays/5:1 Reflux — Approx. of Actual Bolling Point of Components. + TBP — Ambiguous Connotation; Preferable Term Is 15/5 + Stem Correction © — When Thermometers Are Used All the Stem Is Not Immersed. Need to Correct for This. + ASTM Distillations — Handout 1.01 + GC Distillation — Gas Chromatography ("GC's") has Largely Replaced the 15/5. For GC's, the IBP Is Taken as 0.5 Vol.% Point; the FBP Is Defined as the 99.5 Vol.% Point. 1.29 CONVERSION OF ASTM DISTILLATION TO A 15/5 DISTILLATION + Establish Type of Data Reported — Stem Corrected? — Procedure Name? * Plot Data to Arrive at Smooth Curve and Reject Sports * Stom Correct the Smooth Curve, if Not Done by Laboratory + Use Charts in Handout 1.01, Pages 2-3, to Convert This Stem Corrected Curve to 15/5DISTILLATION CONVERSION EXAMPLE ‘Stem Corrected ASTM D-86 (Naphtha) ASTM TEMP Lv% °F 5 347 10 356 50 374 90 401 95 410 FBP 425 50% - 10% = 374 - 356 = 18°F 90% - 50% = 401 - 374 = 27°F 50%-10% 90%-50% increment Increment FO ro -33 20 ant g 16 7 @ Addendum 1.01, Page 4 @ Addendum 1.01, Page 5 PSEUDO - COMPONENTS : Ita Component Analysis Is not Available, a Composition In Terms of "Pseudo" Components Can be Obtained from the Distillation Curve and Gravity. + How Do We Represent the Components? 350 A 45.0 350K 11.6 350 8 0.80 + How Do We Get Pseudo-Components? te 15/5 314 326 374 47PSEUDO-COMPONENT BOILING POINT +H Only an ASTM Dist. Curve is Available, Convert it to a 15/5 Distillation. + Divide the Curve Into Rectangles, Each Rectangle Representing Pesude Component. + Tope of the Rectangle are the Vol. Avg. BP's: 50% of each “Component” Bolla Lower than the Top at Rectangle, 80% Bolle Higher. 010 20-30 40 50 60 70 80 9% 100 Vol. % D +L 12 PSEUDO COMPONENT GRAVITY +Have Tyagp and Vol% of Each Fraction +Have Gravity of Total +Calc. Char. Factor, K, of Total Feed k =_/MeABP, “°R 2 ‘SP. GR. {for Narrow Bolling Fractions, Can Assume Tyapp = Tweasp) + Use Same Equation, Solved for S.G., to Calculate S.G. of Each Pseudocomponent.Page 1 ‘Addendum 1.01 amsseid Bunerade wy 5 ‘1gerj21un fyje10488 eed “10/8u3 wnnven wu OT parie> K\uoWNOD “powIaU UNnDEA DazIpIePUELS A00d “parenbNUY paiayset wnnaen iayjo ay) “sorsagse yuim pajeinsur avo “speay 10den wo10}n1p om, 40 ouo uItM pouiquios (wulgz "4990 Tous 10 “wu OOT "¥290 aden ou1 “pow 2M v) panturad su uoiyy "wowainseau 24, “paUTIsap Aj100¢ pue pat 1uarenindo anrs v0 osn sajdnosowiayy “Seimeroduiey oni jou oie s8uipess s2uay ‘e2eds 10den ut so WASY 1819 0} Ao Kidde y>tus “SIeNBVeW ALII0G-9p1m Pu LnIPOW 14 “W'L°S"y oun Aa pelonb sonten Jo ue! 2) —— — “2a r vonvoyrsads on] x2 | auon | 0st |yatty | 1029) “oy | 090-06 | enpsoy pue| 24 — — — ~—}— | sayeiinsia 5H “Dad soot ost ~ Ot w= =| auoN wor | 2029] “Oy> 0g yori eORLTO : i ~ | T ot )ee-aebuew | sags gas) sods on ---| 2 | auon} ose | vuny | 2845 1°45 | goy-o¢ bux esn)| 851-0 on oN on | on avgjosqg| NISY ean ae a a clarz<| ooee |. ie] 2980 ors 89 ora] ee cz v2] 2 | avon] set | any | zev<]ziz<| o92-0€ |acoeunsia nin] anon z zav<|ziz<| 080 seu2s0r] ors 89 or} 2 Ee wet—— awn] car | ony |= =}————] ‘eden “hny] 2 st zav>|2t2> | 085-0 |sarennsic wor] SY] eereanesliemicries 5 5 “3ad5 | paca, ) AABOH 20 abueyl 298-4 ors 09 os] oz |x v2] st | aon] ser | un | zer>] 25] ogs-oe | “NoHo ciel) | 298-0 HO dad dal] ain do ad siog por eAitjiqovoedoy “1989 4 Pesn sige gerd “PS SOOHLAM NOLLVTTLSIO WASY LN3YUND JO SAUNLVI4 TWdlONINdSTEM-CORRECTED THERMOMETER READINGS FOR ASTM D-86 NAPHTHA DISTILLATION TREADING CORRECTED FOR THERMOMETRIC ERROR; | eer PE SEH Lae Se ae a tae ie at a 7 le ie oe ee ec l oe oo) aie |e gp # #lel|e-e se ee) es 8 b BR |e oR |e ile)e 2 oe gleleg g BR (om fae | Be Be) | se eb | | mw | | mm «© mlm | m om mw om | 2 |e |B # 8 #1 |B R BB me | | i m is @ |i | im bm BR Ble iB BR BB S|a'S BR B 2 Ble 8 8 12/8 8 8 B =» lel me os mw mime | mom a l|g|# 8 @ BIR |B # B B BUS 8 & R BR BS # & B [ela ) eg @ gS Be i Ss |e|e = 2 BIS 2 @ B ee eee ee Ze es ee Ble eg Be B gooey eR FS Blk eg oR eR eee se ee ae eee Be |e B # BRIS # OB vw fel wow woe lw mw Oise 8 of Blaise & oe ee) ge pg Big gs # 8 & Biggie = B Bie |e B BF e|k(S 2 B BI|B|2 & 2 8 40, boa | az a3 a 415 ae ar ae ay 40 sia; 8 # giglg 2 g @ | 2 & |e & & B&B | a mw ieee ce ee se Sean Boe oy fo |S '8)8 8 # S'S) Ss 2 8 8 | 450 | be 465 466 4 as | ae | a0 42 a3 am op te ae a) te eo) ol ce | |# |@/@ £€ & 2 e#\|se #8 8 & 1 ago | a6 | a7 8 500, so | 502 | 503. soa 505, 506 je ksh) Rehr mee ee eee eee ep eee a ie enol ie aoe Se woe |e 2 # Ble 8 # BB Bp) B Bo ele] gos Bt BPE 8 2 #/8)/8 # R R wo [am | me gs oe ais wom |e fijaile 2 # aig og 2 8 elf] 8 8 8 Bie Ao 2/8 & & & e|f 8 @ @ Bg /fi 8 8 8 2/3 82 ¢ 8 wo oa foe taste eet ie alee ase esse ar Page 2 ‘Addendum 1.01Page 3 Addendum 1.01 STEM-CORRECTED THERMOMETER READINGS FOR ASTM GAS OIL DISTILLATIONS, D-86 AND 0-158 saad READING CORRECTED FOR THERMOMETRIC RIOR *F eae (Ghclunte 19 ee OF seed a a ee a a ‘ay ae | ae BR |e es BR BR |e | eo om Re mf | Me | | mM Be ms (om mmm | we 2) S BoM Bo ne mo oom | mo moms m | in _ mm mom | BB mm | mt | BOR | Bo jm lm ROR | m a oo | om | ome aes | a | oma Mf | | bh Oe |B ie Bo me me | mw |e RR | BB Re BR OS | we |e eM Ge ee ee ee wo | am | eae awe awe | ww ee a” eo | BG Be eat ee ee a ee Sf ees | mw | me | mo | we me wm om mw | momma eS | 2G @ & S12 |e 8 2 eee eR Be |e Be Sm | me | em ee op |e me BB | | \ 2 le e w mw eo mw wow & i$ 8 ££ 8 8 oe Bo RM = = © @ © | me me we e 2 2 @€ S & & & so so ee Fw | oe fe ge gts) ge gee Ble) h 2 B RR fF 2 es Bo] | BOBS | Se eh eo w]e | | Se eee Ee i | ee ey eh ee pe ceateectels er (8 @ 2 8 oe | oe | Ss ke Se | Ss | se te et | | ee mt os wm | | a oe a we ww | & | we | te ge Se oe & | & | ee a te et ee a | me | at | te po re) wo lim l im mos am wm wm |) Be ee & 8 & | i | fe ee re ee | | ome mo om hs | % | oR | mom te moO BO me) om lm om om om | m | me momo | BB MM) me | es m | ie | emote os | me | oe Mt m |e | om Sm | me | me mom me BB] RR Room | om | mm om 1 Ling mm oe | m | mm misoF To Be Added To Stem-Comrected ASTM Tenoerature, °F Page 4 Addendum 1.01 DISTILLATION CONVERSION, CORRECTED ASTM 086 OR D158 70 15/5 -20 -10 50 00 SO te 200 eee 250 zea 900) 350 400 ‘Ster-Corrected ASTM 50% Minus 10% Tenveratures, °F 20 20 30 20 50 60 70 Stem-Corrected ASTM 50% Minus 107. Tenperatures, *F°F To Be Added To Sten-Corrected ASTM Temperature, °F EES TEE CORRECTED ASTM 086 OR 0158 TO 15/5 DISTILLATION CONVERSION, 4.00 150 200 Stem-Corrected ASTM 250 300 350 100 Stem-Correctes ASTM 90% Minus 507s Temperatures, °F Page 5 ‘Addendum 1.01Page 6 ‘Addendum 1.01 4+ "404 Sri os o9tt0 MASH ooe___ ow ove ome _—oze_—ooz__—owt’_——gat bt ot oot =o) sommandens O9UTO WAS 8 CL Le ee ee, aC ope rere c2) 44 "Hos Sn Hoe o9TTO MASY S/s¢.01 0919 wisy ayvanvis. or SWSANBOW2ANT NOLL HLSI0 WAROVA ow 02+ ocr ove ost ov ou o9tta WIS¥ 4 300 39 01Page 7 ‘Addendum 1.01“F To Be Added To 15/5 Vacer Tenoeraure Page 8 Addendum 1.01 DISTILLATION CONVERSION, 15/5 To STem-consected [FH [ASTM 086 OR D158 60 50 Ee ao 30 20 ao fy 50 Too 150 200° 250~«*SDSCSC« SO 15/5 50% MINUS 10% Vapor Temperatures, °F 15/5 50% Minus 10% Vasor Terwertures, °FALIAVYDS Idv Ss ° °Q + 30 20 10 40 60 60 70 80 90 100 10 20 30 4 ° Jd - 3univeadWaLALIAVYD Id¥ Ss ° ° oO R o o ° + 20 10 ° o 60 60 70 80 90 100 10 20 30 40 ° ad - 3uNivwadWaL1000 800 700 600 500 8 Temperature - °F _ 8 50 20 30 40 50 60 70 8 9 95 Per Cent Over IGeeem 2aoe4St810) + z,e 1000 700 600 250 Temperature - °F 200 150 T Tt { 50 { 20 30 40 5060 70 80 90 Per Cent Over we 2345 10 987 =PSECTION 2 DISTILLATION GENERAL PROBLEMS (Ref. EEDP—Section III) Maxwell-Chapter 14 1. Design of Fractionation Column to Make Desired Separation of Feed A. Screening B. Definitive Design y - Checking a Given Column for Desired Separation of Feed A. For New Service 8. For Existing Service it We Can Handle the 1st Class, We Will Not Have Trouble With the Second Type. 2ot FRACTIONATION COLUMN (Mole Flow Rates) Condenser Ge Map. Dist. ve Accumulator OHD. Lig. Product ° Xo Enrichi hp Section . f,—feed_| be Low Boling Comp. Sipping Heme, ame.DEFINITIONS 1. Key Components — Major Feed Components Between Which Split Is to be Made. 2, Minimum Reflux — Reflux Required for Desired Separation With an Infinite Number of Plates. 3. Total Reflux — Reflux at Infinite Reflux Ratio. Reflux Required for Separation With Minimum Number of Trays. 4, Minmum Cost Requirement — Cost Optimization for Best Return on Investment. PLATES vs. REFLUX ‘Thore is an economic balance between plates and reflux. Tot. Cost Oper. Cost 10 15 0/D1(01 0) min. Operating Reflux Ratio/Minimum Reflux Ratio |PLATES VS, REFLUX (Continued) * In standard Tower Systems, Economics Not Too Sensitive. Sensitive for Complex System. Can Use Shortcut Procedures Can Use Simplified Plate-to-Plate Calc, * With Utility Cost Going Up, the Optimum. Favors More Trays and Lower Reflux. 205 BASIC STEPS IN COLUMN DESIGN Know— * Feed Rate & Composition * Desired Separation—Recovery & Purity * Battery Limit Conditions on Feed, Products, and Utilities 206STEPS IN COLUMN DESIGN (Continued) 1. Determine Key Components Which Represent the Desired Separation 2. Determine Operating Conditions 3° Use Short Cut Distillaton Techniques to Define the Extremes—Min. Stages & Min. Reflux 4° Select an Operating Reflux Ratio 5." Use Gilliland or Similar Correlation to Determine No. of Theoretical Stages 6. Pertorm Plate-To-Plate Calc. on Computer to Define the Tower 7. Optimize Tower Design —Reflux —Number of Stages —Level of Feed Preheat —Feed Tray Location 207 INITIALIZING COMPUTER PLATE-TO-PLATE CALCULATIONS TO MODEL THE TOWER * Steps 3 to 5 May Be Eliminated by Use of Information From a Similar Operations PROCESS—COLUMN Has a Short Cut Distillation Procedure to Help Initialize a More Detailed Plate-To-Plate Tower Simulation, (Uses a Generalized Fenske Method) * Addendum 2.2 Provides More Details on Short Cut Distillation Techniques 208OPERATING PRESSURE FACTORS TO CONSIDER 1, Disposition of the Overhead Product 2, Is the Overhead to be Totally Condensed? 3. Temperature of Available Cooling Medium to Overhead Condensers. It Cooling Water Available at 95°F is Used, the Min. Pressure for BP Distillate Would be That Corresponding to the Pressure at a BP Temperature of About 110°F (15° Approach). 209 OPERATING PRESSURE (Continued) 4. The Max. Pressure Might be Limited by A. Tower Bottoms Temperature. A Very High Pressure Could Result In Excessive Bottoms Temp. (Cracking, Color) B. Temperature of Heating Medium for Reboller C. A Bottoms Temperature too Close to the Critical Temperature Results in Poor Separation—Bottoms Temperature Should be at Least 50°F Below T, 210OPERATING TEMPERATURE. 1. Fixed Once the Separation and Pressure are Specified. 2. Temperature Estimates. A. BP Calculation on OHD at Accumulator Pressure Gives Reflux Temperature for Total Condenser. Flash Calculation lf Not a Total Condenser, 8B. OP Calculation on Tower Overhead Will Give Tower Overhead Temperature. C. BP Calculation on Bottoms Will Give Bottom Temperature. 0. Flash Calculation on Feed to Get Temperature & Fraction Vaporized. 2m FEED CONDITION Factors to Consider— 1. Balancing of Tower Loadings to Take Advantage of Avaliable Column Diameter Above and Below the Feed 2. Cost of Column Reflux (Water or Refrigeration) Cost of Column Rebolling 4. Heat Available From Exchange With Products or Overhead Vapor 5. Not Unreasonable to Start With BP Liquid 2 222RELATIVE VOLATILITY (USED TO MODEL SEPARATION) Comp. Ky a= KyKe A Ky KaiKe 8 Ka KelKe Heavy Key —> c Ke 10 Component D Ko KpKe Ky. Pale Vapor Pressure Component A = Ka 2 Palt . py jp, — Vapor Prossure Component “Ac = Ko = Bein * °&!Pe Vapor Pressure Component C MINIMUM REFLUX (UNDERWOOD’S EQN) =N xix, O A, (Baum be x in6 = ( “oan oat = Min. Reflux Ratio Where: Xi < @ < Ox f=N iz, « it i Moles Reflux (O/D) MIN = Minimum Reflux = jouer Mole Fraction Liquid in Feed Moles Fraction I in Distillate. Mole Fraction | in Feed Relative Volatility of i (ip Xp Xp)" + Geometric Mean Average 214MINIMUM STAGES (FENSKE EQN) ( (% Uses: onl G2) Ge2)] Estimate Minimum No. of MN Trays Required —Estimate Feed Tray Location Sy = Minimum Number of Stages (At Total Reflux) Xixo = Mole Fraction Light Key in Distillate Xixe = Mole Fraction Light Key in Bottoms Xo = Mole Fraction Heavy Key In Distillate Xyxe= Mole Fraction Heavy Key in Bottoms Relative Volatility of Light Key (Averaged Over the Entire Tower) = (Xing X Cine x Xixy)"> = Geometric Mean Average t t t Tower Tower Tower Bottoms Feed —_—Distita 215 ESTIMATING ACTUAL REFLUX RATE AND ACTUAL NUMBER OF STAGES General “Rule of Thumb"— (3) cua = (1.1810 128) x (Baan Use Gilliland Correlation to Calculate Actual Theoretical Stages for Selected O/D (P, 244 Maxwell). Use Safety Factor of 10% or 3 Trays, Whichever is Larger. 216FEED TRAY LOCATION 1, Locete so That Ratio of Key Components in Feed Liquid is About Same as Liquid on Feed Tray. (Rule of Thumb) 2, Compare With Plate-To-Plate Printouts of Similar System 3. Use of Fenske EQN. at Total Reflux 4, Parametric Study on Plate-To-Plate Calculation (e.q. PROCESS-COLUMN) or Parametric Study Using a “Short Cut” Distillation Tower Model (e.g. PROCESS-SHORTCUT) 2a FEED TRAY LOCATION (MODIFIED FENSKE EQN.) ‘Approximate Feed Tray Location Set by the Ratlo— Sap Su Number of Stages Above Feed (At Total Reflux) Mole Fraction Light Key In Distillate Mole Fraction Light Key In Feed Xtxp = Mole Fraction Heavy Key In Distillate Xyxr = Mole Fraction Heavy Key in Feed o'x = Relative Volatility of Light Key (Averaged Over the Enriching Section of the Tower) = (euxeX ouxo)2 4 Tower Tower Feed Distillate ateFEED TRAY LOCATION (KIRKBRIDE METHOD) This Is the Correlation Used by the Shortcut Distillation Technique in PROCESS/PRO II Sar [ Kane) (Xuxe ’] og =F = 0.206 log | Bip (uxe| 109 Spp = 0208 log [Pi ie) Xan) Sarigg, = (Ky)?20 where K, = [Bip (ZHt*) Xua)'] Xue’ \Xako COMPARISON OF KIRKBRIDE METHOD WITH FENSKE METHOD: (MOLE FLOW RATES AND COMPOSITIONS) XicrH0.25 Oy, Kore 7025 Sic =148 (Entire Tower) Xixo 20.60 Xixg=0.03 xp =OL LK (Enriching Section) Xj =0.03 Xjxa=0.10 Bip=153 FENSKE METHOD KIRKBRIDE METHOD S=12.22 Sap=3.42 Sar o Ser=Sy-S,e=8.80 logg;, 70.206 logl(.63\(0.20y1.0)"] Sarj = 0-78 Spr 220TRAY CHARACTERISTICS (REF. EEDP—SECTION III-A, TABLE 1) Type Capacity Efficiency __Cost Flex. Sieve Med.to High Lowest Med, to Good High ee ON eee BC. Low to Med. to High Med. to Good Med. High 200% Sieve 3/1 Valve Goodas Good as Med. 10- High Up Sleve. Sieve 20% >Sieve to 5/t Jet Highest-HP Lowto 5%>Sieve Low. 1.5/1 or Med-LP = Med. 2n 221 SINGLE PASS Dc. Clearance Liquid Sieve Tray (Shut Down)FR TWO-PASS TRAY Weir Height “Tot pod “T “oat TRAY NOMENCLATURE — SINGLE PASS TRAYS sLOPED DOWNCOMERS SHOWN) fae NOMENCLATURE. Moa Mat he ma, A, = Free (vet Aron = Beaten to det lood Point = Bubble Area = Tower Cross Sectional Area rrp ‘Ag = Down Gomer Inlet Area ‘Ay = Down Comer Outlet Area AL = Waste AreaTRAY NOMENCLATURE — TWO PASS TRAYS (CHORDAL DOWNCOMERS SHOWN) “ oe il ya rhode a Lie Aa : ore fersuneore en Cc a w | For Design Caics \ Use the Smaiier eee aa, | ofthe hwo" + = inboard ey PRESSURE BALANCE FOR A TWO PASS SIEVE Pressure Balance for Inboard Downcomer Filling: Lg = Pe + PAH Lhe Pressure Balance for Outboard Downcomer Filling: Lg = Pag + Py tb be KS Total Tray Pressure DropTYPICAL TRAY PERFORMANCE DIAGRAM Poor Ett. Jet Flooding Below Design ticleney Increasing Vap. Rate—= Dumping Increasing Liq. Rate—= TOWER SIZING For Sieve Trays A. Calc. a Trial Tower Diameter With Assumed Tray Spacing . Check Liquid Carrying Capcity Repeat (A) if Necessary Size D.C. & Set Other Tray Details . Calculate Tray Hydraulics Calculate Perforated Area . Calculate Flexibility |. Calculate Efficiency ZO™moo® 228COMPUTER PROGRAM SELECTION FOR PLATE-TO-PLATE CALCULATIONS * PROCESS-COLUMN or COPE-TOWER — Good for Simple and Complex Systems — Direct Access to EDL — Can Set Product Specs — Can Run With Minimum Data — Maybe Difficult to Converge With Complex Systems Depending on Nature of the Complex System, Either TOWER or COLUMN May be Preferred ASPECT Il —Fast — Can Use Steam Stripping — Wider Range of Temperature and Pressure — Good for Heavy Hydrocarbon Systems (Vacuum Pipe Stills) * CHEMDIST — For Severely Non-ideal Systems — Limited in Number of Components USE OF PACKINGS Vacuum Systems Where Low AP is Critical Debottlenecking Existing Trayed Towers Light Ends Towers at High Liquid Loadings Absorption and Stripping Services (High Liquid Loadings) Small Diameter Towers (Tower Diameters Under About Three Feet) Corrosive Services Where Ceramic Materials or Plastic Materials are Attractive. Corrosive Services Where Packing Cost is Less Than Alloy Tray Cost 220 229PACKING CHARACTERISTICS AND TYPES (REF. EE. DP-SECTION III-A, TABLE 2) Type Capacity _Etticlency Cost Flex. indom Dumped Packing (eg. Pall Rings, Metal Intallox, Nutter Rings, ete.) Structured Packing Med, to. Med. to High (at Med. to Good (eg. Flexipac, Montz, Very High Very High Least 2x > 3 Gempak, etc.) ROP Cost) Grids (eg. Giltsch Very High Poor (Good = Med. to. = Low < 24, Grid, Flexigrid, For Entrain- High ‘Snapgrid) ment Removal and Heat Transfer) Based on Efficiency, Capacity, Tumdown and Cost, the Two Inch Metal Pall Ring Is an Economic Packing for Most General Petroleum Distillation Services. 231ADDENDUM 2.1 Section 2 DISTILLATION Nomenclature = Area, ft? = Bubble area, ft? = Downcomer clearance area, in? Downcomer inlet area, ft? = Downcomer outlet area, ft? = Average tower free area, ft” (superficial area minus arithmetic aver- age of inlet and outlet areas of downcomer (s) above the tray minus the waste area); for multipass trays, use smallest value of Ay. = Hole area, ft = Superficial (total) tower area, ft? Waste area, ft? (normally zero for sieve trays) = Clearance between tray and downcomer at tray inlet, inches Diameter, ft Trial diameter, ft Overall efficiency, % Peeper " rMg0°PP>P " Tray spacing, ft he = Clear liquid height on tray, inches of hot liquid = Downcomer filling, inches of hot liquid heq = Effective dry tray pressure drop, inches of hot liquid hy = Try inlet head, inches of hot liquid hy = Total tray pressure drop, inches of hot liquid tug = Head loss under downcomer, inches of hot liquid hy = Inlet weir height, inches two = Outlet weir height, inches Ky. = Tray spacing—liquid rate capacity factor, dimensionless Kyp = Liquid height—-pressure drop factor, dimensionless Koz = Surface tension—viscosity capacity factor, dimensionless Kjp = Hole diameter—pressure drop factor, dimensionless Ky = Clear liquid height—density difference weeping factor, dimensionless Ke = Tray spacing—percent hole area weeping factor, dimensionless Kg = Surface tension—hole diameter weeping factor, dimensionless L = Liquid rate, gphitt of weir per pass u = Liquid rate, gphitt of diameter per pass & = Liquid load, ft°/s at conditions Lug) = Minimum liquid load, t%s at conditions mmosaon0 Page 1Nomenclature (Continued) f 4 6 fa M M, Np Q Vai vr Me Vesti Vo Voguin) Vocweep) A AL Pv Oo Isto jossoazon Flow path length (distance between inlet and outlet downcomers), feet Inlet weir length, inches Outlet weir length, inches = Length of bottom edge of downcomer, inches = Liquid loading, Ib/s Vapor loading, Ib/s = Number of liquid passes Liquid rate, gpm at conditions Allowable velocity of clear liquid in downcomer inlet, ft Vapor velocity based on average tower free area, ft/s 8 Design vapor load = ® Ve at conditions Ultimate vapor load dependent on system properties, f°/s at conditions Vapor velocity through holes, ftis Vapor velocity through holes at minimum vapor rate, fts Calculated vapor velocity through holes at which tray will start to weep, fts Capacity factor, TN a Liquid viscosity at conditions, cP Liquid density at conditions, Ib/tt? Vapor density at conditions, Ib/ft? Liquid surface tension at conditions, mN/m Standard liquid surface tension, mN/m oe) 10 ww Hole diameter, inches (160 - Page 2ADDENDUM 2.2 Section 2 DISTILLATION In order to simplify the work involved in making stepwise calculations for the rectification of bi- nary and multicomponent systems, Gilliland’ has presented an empirical correlation between theoretical steps and reflux ratio. To use the Gilliland correlation to predict the number of theo- retical plates for a given reflux ratio, the minimum number of steps at total reflux and the mini- mum reflux ratio are required. Minimum Number of Theoretical Steps When a separation is specified with respect to only two components of a multi-component mix- ture, the lower boiling of these two components is designated the light key component and the higher boiling the heavy key component, and the minimum number of steps can be calculated by the well-known Fenske equation? as follows:? lolita) Gt) Su = Tog ak ) or x, lau = (te) a (ra) After equation (1) is solved for Sy, the latter may be substituted in this equation along with the distribution of either key component to predict* the distribution of the other components, or tog (42) tt) = Sy tog @ Likewise, tog ($E24-)(RU) = sy tog (4) ® In any of the above equations, moles per 100 moles of feed may be replaced by total moles, or volume or weight units since in any of these conversions the multiplying factors cancel out. When the degree of separation is specified for more than two components, equation (1) must be applied to all critical combinations of these components and the maximum Sy determined for the most difficult case. If the separation is specified with respect to the total quantity of two ‘of more components, as in the case of Example 1, trial and error is required for the solution of Su. It should be pointed out that the concentrations calculated by equations (2) and (3) actually ap- ply only to the separation at total reflux and, with the exception of the two key components, there will be some variation of the degree of separation with the reflux ratio, As the reflux ratio decreases, there is some improvement in separation between light and heavy components boiling outside the range of the key components and some deterioration in the separation of components boil- ing intermediate between the key components. However, in so far as the present procedure is concerned, the distillate and bottoms compositions for other reflux ratios are assumed to be the same as those calculated for total reflux. ‘Gilliland, ind. Eng. Chem. 32, 1220 (1940). 2A table of nomenclature is given on page 10. *Fenske, Ind. Eng. Chem. 24, 482 (1932). “This equation may be used for any pair of components. cecoezere Page 1Minimum Reflux Ratios Gilliland® has proposed several different formulas for predicting minimum reflux ratio and all have the disadvantage of being composed of a number of complex terms in addition to requir- e ing trial and error for solution. Although all these equations appear to give satisfactory results, ‘the terms are so complex that it is difficult to be certain that there are no numerical errors in their application. In order to apply the Gilliland method with greater facility, the following equation was developed for predicting the minimum reflux ratio of a multicomponent system: (Oy + + = (ae ~Xino) - Pac ani Qin ~ WXxxo) + Faxon meee - 20) @ (O/D)y can be calculated for two arbitrary states of feed vaporization: 1. “‘Liquid” feed, corresponding to vaporization of the feed equivalent to the fraction of the feed lighter than the light key component. For the components lighter than the light key, I. = Zi/ay, and for the light key and heavier components, lx = Zix, and Iy = Zy.° 2. “Vapor” feed, corresponding to vaporization of the feed equivalent to the fraction of the feed Consisting of the heavy key component and lighter. For the components lighter than the heavy key, IL = Zila and Ix = Z,x/c%,« and for the components heavier than the heavy key, ly = Zy After the minimum reflux ratios have been calculated for “liquid” and “vapor feeds, the mini- mum refiux ratio for the actual vaporization of the feed can be calculated by direct interpolation e or extrapolation, However, extrapolation beyond 50% of the difference between “liquid” and “vapor” feed may lead to serious deviations. The first term of the right-hand side of equation (4) is the same as for binary mixtures, and the equation reduces to the equivalent of a binary mixture when alll light components other than the light key have infinite volatility and all heavy components other than the heavy key have zero volatility. Under these circumstances the equation is exact when I, is taken as the ratio of the two components in the liquid phase of the feed. That is, if the feed is introduced as a liquid at its bubble point, Ix = Ziq, which is the ratio of the two components in the feed, if the feed is introduced as a vapor at its dewpoint, I, = Z,x/o1,x, which is the ratio of two com- ponents in the equilibrium liquid, For intermediate stages of vaporization I, can be calculated from the flash vaporization formula, although direct interpolation of the minimum reflux ratio of the basis of percentage vaporization between the saturated liquid and saturated vapor feeds gives values only slightly in error on the conservative side. In the case of multicomponent mixtures, equation (4) is semi-empirical since it was necessary to make simplifing approximations in its derivation. Furthermore, the exact values of the various I's cannot be calculated directly from the composition and state of vaporization of the feed, since the liquid on the feed piate is not identical to the liquid phase of the feed as in the case of a binary mixture. As a result, it was necessary to define the I's empirically for two states of feed vaporization, arbitrarily chosen to simulate a binary mixture of the two key components, and then interpolate or extrapolate to the minimum reflux ratio corresponding to the actual vaporiza- tion of the feed. e SGilliland, Ind. Eng. Chem. 32, 1001 (1940). Sif components, intermediate between the two key components, are present, they are considered either light or heavy components depending upon which key their volatility more nearly approaches. In the case of “liquid” feed, I, = Z, and ly = Zy for these intermediate components; in the case of “vapor” feed, IL = Z,/ay and hy = Zyoin soso Page 2Equation (4) has been checked for a number of multicomponent systems on which the mini- mum reflux ratio was determined by stepwise trial and error calculations. Generally, unusual systems were chosen with respect to composition and relative volatility in order to reveal the maximum deviations ever likely to be encountered in practice. The agreement was quite satis- factory as the average deviation was less than +5% and the maximum about 10%. The latter occurred at the limit of extrapolation relative to the arbitrary feed states. Also, the minimum reflux ratio was calculated for these same systems by the Colburn method” with about the same degree of accuracy. It should be pointed out that the latter gave better results than equation (4} when the relative volatilities and compositions were not so abnormal as the system selected. However, under these circumstances both methods were quite accurate as the deviations seldom exceeded a few percent, and the present equation has a distinct ad- vantage in that it is explicit and does not require trial and error. Both methods are quite sensitive to the selection of key components, and the selection of the wrong key components can lead to a much greater error than is inherent in either method. If the desired separation is between adjacent components, there is usually no doubt about select- ing these as the key components. However, if there are additional specifications relative to other components, it may be necessary to try two or more combinations of key components to make sure that the minimum reflux ratio is sufficient to fulfull all specified conditions. Correlation of Theoretical Steps With Reflux Ratio As mentioned at the beginning of this section, Gilliland correlated the results of a large number of stepwise calculations on various binary and multicomponent mixtures by plotting [IS - Sy] /[S + 1] = $(S) against [((O/D) - (O/D)y] / [O/D + 1] = F(O/D) and found that all points could be represented by a single curve irrespective of the type or degree of separation. These points, along with about half again as many additional points, were replot- ted, and the best curve through them was essentially the same as Gilliland’s original correlation. In arriving at the coordinates for the additional points the minimum reflux ratio was calculated by equation (4); therefore these points are a criterion of the present method as well as the curve itself. In no case did the deviations exceed either 3 theoretical steps or 15%, and the average deviation was less than 1 theoretical step and also less that +5%. To take care of the maximum deviation it is recommended that in any design the number of theoretical steps predicted from the correlation on page 13 be increased by either 3 theoretical steps or 10%, whichever is greater. Plate Efficiency Because of the large number of factors which undoubtedly influence the plate efficiency of a fractionating tower, any fundamental formula accounting for even the most important variables must necessarily be quite involved. For this reason, a simple empirical correlation of the limited data on hydrocarbon mixtures seemed to be the most promising method of predicting plate efficiency. Gunness® correlated the results of several tests on petroleum mixtures on the basis of vapor pressure of the liquid. As he points out, this is a method of indirectly correlating plate efficiency with liquid viscosity since viscosity of pure hydrocarbons and narrow boiling fractions is an ap- proximate function of vapor pressure over a fairly wide range of vapor pressures. 7Colbum, Trans. Am. Inst. Chem. Engrs. 37, 805 (1941). ®Gunness, Sc.D. Thesis, Mass, Inst. Tech. (1936). oasoazone Page 3In view of the consistent results obtained by Gunness, plate efficiency was plotted directly against fluidity (reciprocal viscosity) for a number of tests on commercial towers including those upon which Gunness based his curve. The curve on page 12 represents this correlation. While the over-all plate efficiency exceeds 100% at fluidities greater than 9 cm™", this is not inconsistent as the flow of the liquid across the plates results in concentration gradients which may achieve a greater degree of fractionation than predicted by stepwise calculations in which the liquid is assumed to leave the plate in equilibrium with the composite vapor. Lewis® has shown theoret- ically that different combinations of liquid and vapor concentration gradients across the plate may give over-all plate efficiencies as high as 200-300% when based on stepwise calculations. There is no reason to believe that this correlation applies to mixtures other than hydrocarbons, and with the exception of alcohol-water mixtures there are too little data available to afford a comparison. Although there is considerable variation in the alcohol-water data, there is some indication that plate efficiencies are somewhat greater than for hydrocarbons of the same viscosity. Location of the Feed Plate As a simple approximation for locating the feed plate, it may be assumed that the proportion of actual plates above the feed will be the same as that required to effect the same separation between the key components at total reflux. That is, the number of theoretical steps at total reflux is calculated for the concentration then assumed that the ratio of this to the total number of theoretical steps at an infinite reflux ratio is the same as the ratio of actual plates above the feed is to the total number of plates. Application of this method is illustrated by Example 1. In some cases where there are critical components other than the two key components, it may be necessary to check the total reflux steps above and below the feed on the basis of these components, since the optimum location of the feed plate will be different with each pair of com- ponents. Usually the separation of components other than the key components is so complete that only the latter need be considered. Example At an operating pressure of 100 psig determine the number of plates and reflux ratio required to separate the mixture given below so that the bottoms contain at least 90% of the butenes-2 present in the feed and at the same time have an isobutene content not greater than 5%: Feed Component (Mole %) i-CgHi0 40.0 iC,He 200 CyHert 15.0 CaH10 50 t-CyHyr2 100 o-CyHy2 10.0 100.0 Lewis, Ind. Eng. Chem. 28, 399 (1936). ooooszo Page 4(1) Dewpoint of Distillate and Bubble Point of Bottoms In order to calculate the average volatilities, the dewpoint of the distillate and bubble point of the bottoms must be found by trial and error using assumed compositions. These are tabu- lated below. co Moles Per 100 Moles of Feed Mole Fraction an Feed Distillate [ Bottoms | Distillate | Bottoms +-G4Hio 400 293. =| OF (0530 0027 iCjHg 200 187 13 253 050 CaHet 150 130 20 176 or7 CoH 50 10 40 014 184 tO,Hr2 100 15 a5 020 327 o-CyHg-2 100 05 95 007 365 7000 74.0 260 7.000 7.000 As a first trial, assume the dewpoint of the distillate is 140°F at 78 atm (114.7 psia). First Tal ‘Component Yo a - i 140°F 140°F Tye -CuHio 0530 1.29 a4 0.493 FC yHy 253 1.155 75 263 Cae 176 1.13 735 187 CaHio 014 1.00 65 017 tOyHe2 020 os7 63 025 oC He2 007 ost 59 009 7.000 0994 Relative volatilities to C4H yo or (c's aro used as a matter of convenience; then, the (a’'g,)'s are converted to (cqy)'S, the relative volatilities to t-C,Hg-2, which will be selected as the heavy key ‘component. **Computed from the fugacity function of butane multiplied by the relative volatilies. Since the sum of the x's is 0.994 instead of 1.000, the assumed temperature should be lower slightly, but the difference would be so small (less than 1°F) that the change in the (x‘p)'s would be imperceptible. Consequently, 140°F will be used as the dewpoint of the distillate. p Page 5The bubble point of the bottoms is assumed to be 165°F at &0 atm" for the first trial. First Trial Second Trail Component] Xy awe FY y awe Pe tesr | 165°F | Fem | 160°F | 160° y iCHi | 0.027 126 | 107 | 0036 1.265 | 10.25 | 0035 i-C4Hy 050 144 97 061 1.145 93 058 CyHet 077 1115 95 091 112 94 088 Cio | 154 1.00 a5 164 1.00 at 156 tCHye2 | 327 og7 325 | 337 097 785 | 321 CC He2 | _365 oss | 78 356 oss | 74 338 1.000 1.045 0.996 "Relative volatilities to CgHo or (cr)'s are used as a matter of convenience; then, the (%'gy)'s are converted to (%ay)'s, the relative volatilities to t-C,Hg-2, which will be selected as the heavy key component. **Computed from the fugacity function of butane multiplied by the relative volatilies. The bubble point of the bottoms will be taken as 160°F. The relative volatilities are averaged and converted to t-C,Hg-2 as the heavy key in the following table: c 1 | 1408 ‘eo 1500 Oy -omponent °F PF pone ee a, 78 atm 80 atm 79 atm _|(2'oa!wa'a) 7 -CyHig 1.29 1.265 1.275 1.275 1.315 i-C,Hg 1.155 1.145 1.15 1.15 1.185, CyHp-1 113 1.12 1.125 1.125 1.16 Caio 1.00 1.00 1.00 1.00 103 Cater? og7 os7 os7 os7 100 c-C,He-2 0.91 0915 091 0.91 0.94 (2) Minimum Theoretical Steps (Total Reflux) The minimum number of theoretical steps by which the desired separation can be accom- plished is calculated as follows: Let t = moles of t-C,H,-2 in the distillate per 100 moles of feed 10 — t = moles of t-C,Hg:2 in the bottoms per 100 moles of feed Since 90% of the butenes-2 must be retained in the bottoms, the cis-butene-2 content of the distillate and bottoms will be: and (2 — t) moles in the distillate per 100 moles of feed (@ + 1) moles in the bottoms per 100 moles of feed "After allowing 3 Ib/sq in. as the approximate pressure drop through the tower. Page 6Using the previously assumed values of 18.7 moles of isobutene in the distillate and 1.3 moles in the bottoms, the following equations must be satisfied: (8) (a) «ses , (8) (ge) = (ay A trial and error solution of these equations shows that they are satisfied by Sy = 255 andt = 1.62. The distribution of the other components can be calculated from Sy and the distribution of tC,Hy-2. t-C,Hio: Let u = moles of t-C4Hjo in bottoms (4) ) (238) = (raters = 1075 = 0.19 moles of t-C,Hio in the bottoms CyHe-t: Let v = moles of C,Hg-1 in the bottoms v) (8.38 255 C3) (FR) = ase = 44 V = 158 moles of C,Hg-2 in the bottoms CaHio: Let w = moles of CyHyo in the bottoms (S—w\ (8.38) _ 255 _ ( ) (838) = (1.03) = 2.12 W = 355 moles of C,Hio in the bottoms The percentage of -C,Hg in the bottoms will be: 13 (onaneeisree) = 100 = 53% In order to meet a maximum of 5.0% t-C,Hg specified for the bottoms, it is necessary to reduce the 1.3 moles to 1.22 moles in the bottoms. This would require an increase in Sy to 25.8 which would modify the distribution of the other components. However, the latter change is so slight that it can be neglected. The composition of the overhead and bottoms will then be: mt Moles Per 100 Moles of Feed Mole Fraction Se Feed Distillate | Bottoms | Distillate | Bottoms iC aHio 400 3981 0.19 0528 0.008 i-CyHy 200 1878 1.22 229 050 CyHe-t 150 1342 158 178 064 Caio 50 1.45 355 019 145 tCHy2 100 162 838 021 342 c-CyHs-2 100 038 62 005 391 75.46 2454 4.000 4.000(3) Minimum Reflux Ratio ‘Since the critical separation is between isobutene and the butenes-2, the former is naturally selected as the light key component and trans-butene-2, since it is more volatile than the cis-butene-2, as the heavy key component. Butene-1 is considered a light intermediate com- ponent because of the proximity of its relative volatility to that of isobutene; normal butane is considered a heavy intermediate component since its relative volatility is nearer to the heavy key than the light key. The following tabulation gives the necessary information for calculating the minimum reflux ratios for the two arbitrary states of feed vaporization: Mole Fraction 1 Component] Type = ay | “Liquid” | “Vapor” Feed | Distillate | Bottoms eat | feed FOHio L a400 | 0528 | 0008 1315 | 304 304 1-C,He uk .200 249 050 1185 | 2.00 169 CHet | ob 150 178 064 1.16 150 1.29 Cato H 050 019 145 1.03 400 348 tC,Hy2 | HK 100 021 342 1.00 = = eCyHy2 | H 100 005 391 094 2.00 2.00 1.000 | 1000 | 1.000 “Liquid” feed—40% vaporized (1D + 1 = (13955200510) + (438) (0528 - 304 x 0021) + 448 (0.178 - 150 x 0021) + sabe (BP - oor) + reBBbur (BBP - 008) (OlD)y = 188 + 194 + 107 + 0.29 + 046 - 1 = 464 “Vapor” feed—90% vaporized (O1Dyy + 1 = MBSR LO (0289 _ ooo1) + 194 + $48 (0.178 - 1.29 x 0021) + tres CRB - o019) + 046 (Oly = 206 + 194 + 1.40 + 035 + 046 - 1 = 491 ‘Assume that the feed is sufficiently preheated to vaporize a percentage equivalent to the distillate or 75.46%. By interpolation, the minimum reflux ratio corresponding to this feed vaporization is: (O1D)y = 464 + (7848549) (401 — 464) = 483 Slama Page 8(4) Theoretical Steps vs Reflux Ratio Using the values determined in preceding sections for minimum theoretical steps, 258, and for minimum reflux ratio, 4.83, the number of theoretical steps for various reflux ratios can @ be predicted trom the correlation on page 13. F(O/D) (8) s [Theoretical Plates" = = o 5 0.067 0570 613 603 136 502 52.7 51.7 223 430 460 45.0 314 366 413 403 - _ 258 248 *The reboiler is considered the equivalent of one theoretical step. With a partial instead of a total condenser, a second theoretical step also could have been deducted. (6) Number of Actual Fractionating Plates To predict the number of actual plates it is necessary to determine the average viscosity of the liquid on the plates. Since the temperature difference between the top and bottom of the tower is so small, the average viscosity may be taken as the viscosity at the average temperature. For this purpose the viscosity of butane at 150°F will be used. Viscosity of CyHyo @ 150°F = 0.216 cs ~ 0.216 x 0523 = 0.113 cp Fluidity = 10.113 = 89 cp~'; Plate efficiency = 99% e Using a plate efficiency of 99% the number of actual plates is computed for various reflux ratios: ) s Theoretical Steps | Actual Plates 483 oo oo oo 5.25 613 603 609 575 527 517 52.2 650 460 450 455 750 413 403 40.7 oo 258 248 25.0 The number of actual plates is plotted against reflux ratio in Figure 1. Areflux ratio of 6.50 to 1, or 1.35 times the minimum is selected. The number of actual plates corresponding to this reflux ratio is 45.5 so that a 50-plate tower would be required. (©) Location of the Feed Plate ‘The number of plates above the feed is based on the proportion of theoretical steps at total reflux which would be required to effect the change in concentration of the key com- ponents between the feed and distillate. This proportion is applied to the actual number of plates (including the reboiler) to determine the number above the feed plate. In order to take into account any appreciable difference in relative volatility above and e below the feed, the relative volatility used for calculating the steps at total reflux between feed and distillate is the geometric mean of aip and cg OF, on = (HEP x 889)" 110 osooszaa0 Page 9‘The number of total reflux steps which would be required between the feed and distillate is calculated by the following equation: (3) () = 1.19" = 5.78;n = 10.1 Number of actual plates above the feed would then be: BA 0 + 1) = 20 The vaporization of the feed can be taken into account by adding the fraction vaporized to n since 100% vaporization would be equivalent to a theoretical ‘step at total reflux. This would change the proportion of plates above the feed as follows: (144975) (5041) = 21.4 plates above the feed Feed lines would probably be installed above the 24th, the 28th and 32nd plates from the bottom of the tower. Nomenclature X moles of any component in distillate or bottoms per 100 moles of feed x mole fraction of any component in liquid y mole fraction of any component in vapor D _ moles of distillate per 100 moles of feed O moles of reflux per 100 moles of fee o/D retlux ratio (O/D)y, minimum reflux ratio corresponding to S = oo S "number of steps from still to distillate Sq minimum number of steps corresponding to O/D = oo P number of theoretical plates; with a partial reboiler and partial condenser, P = S ~ 2, and with a partial reboiler and total condenser, P = S — 1 Zz atio of mole fraction of any light component to heavy key component in the feed Z4 —_Tatio of mole fraction of light key component to any heavy component in feed 2 relative volatility of any component to heavy key at the dew point of the distillate yw relative volatility of any component at the bubble point of the bottoms %, _relative volatility of any component at the arithmetic average temperature of the dew Point of the distillate and the bubble point of the bottoms gy mean relative volatility of any component, (ap-aw - a4)“ LK used as a subscript to refer to the light key component HK used as a subscript to refer to the heavy key component L used as a subscript to refer to any light component H used as a subscript to refer to any heavy component D used as a subscript to refer to the distillate W = _used as a subscript to refer to the bottoms n used as a subscript to refer to the Plates above the feed m —_used as a subscript to refer to the plates below the feed a Pana 10FIGURE 1 = 25 Actual Plates 3 5 3 E 3 z 5 £ = = e8'y = ney xnjey WNWTUIW ‘Sa1e]1q feNoy JO JequiNN Reflux Ratio 60; 3 $ + ry + ° Page 11Overall Plate Efficiency — % 120, 0 FIGURE 2 OVERALL PLATE EFFICIENCY VS FLUIDITY OF LIQUID ON PLATES Only Data on Hydrocarbon Mixtures Were Used in this Correlation, and There Were Insufficient Data on Other ‘Types to Justify a More General Use. However, There was Some Evidence that the Curve is a Little Conservative for Alcohol-Water Mixtures 1 23°45 67 8 9 10 11 12 13 14 Average Fluidity of Liquid on Plates — (Centipoise)-' Page 12FIGURE 3 GILLILAND CORRELATION CORRELATION OF THEORETICAL STEPS WITH REFLUX RATIO MULTICOMPONENT AND BINARY MIXTURES ub & Roo NN ® bo Fe > Reference: Gilliland, Ind. Eng. Chem. 32, 1220 (1940) R-Rm, R F(R)+Rm FOID) = FR) = "Ray Am * Rmft-FR)] s-sm, S (S)+Sm OS) = ~SEr Sm = Smlt- GO @ &(S) ub oa ON & © 12345678 910 F(R) or F(O/D) Page 13accrion(ai FRACTIONATING TOWERS DISTILLATION DEVICE SELECTION AND BASIC CONCEPTS e TABLE 1 TRAYS - A SUMMARY OF CHARACTERISTICS [rare a = mane | ees | mae Bee. ee Weim | ponte ™ data avedadie. wa a aaa | coe Seen [aT Soom so Saget me arias — Taare ae | ee, Same fees [Sars [sor [Se [Sat ae Sure, | Sen See ace = ae em ae Semon =a tee |e mca ie ead See - ae meer PCR tao Ge irene a een ona eee =F a ce et =7s. Tore |S Peer ia STs a eR Rg ee o Eanom aon sme ae an Sor, = soa LL @ TABLE 2 COUNTER-CURRENT DEVICES "A SUMMARY OF CHARACTERISTICS Seer — ase 1 eisoey tat rant res Tes? Te = ro a wpe ao Ss ae an. =. SS ts = reas = rr a a a sooo Merce” [Saacst [Mnasst | ron‘ =e ee. [oo [so er = Soe ae ES oro ee ae = poses Ea Seams = re Pte) roa i= x GoTo ae Sees | oe ae ae = am aon eae = was wiares one Sacto « Pins on ln cane See [Rae Sa" |sre Sane m a ax =e, pS sa = a TA eR SRT Ee SRST HT Raa ae Ta be | |SODIAIOS Joysuely Jeay ‘Burjog ‘Hbuljno4 - Aouaioyjy Mo ‘Ayoedes ybipy, + puBdeus seynNDESIGN PRACTICES EXON ‘Section Pope FRACTIONATING TOWERS me 40 of 62 PACKING AND GRID PROPRIETARY INFORMATION — For Authorized Company Use Only EXXON ENGINEERING Figure 4 ANCILLIARY TOWER INTERNALS. NEEDED IN PACKING INSTALLATIONS ‘Vapor Out VL wus" (Only If Deentrainment Required) Remove Existing ~~) Tray Supports and (To Within 4" to 4” e [6-13 mm] of Wall) Gas Injection ‘Support Plate Liquid Product onmney oY or Drawott (For Sidestreams, Pumparounds, etc.) Liquid Distributor (For All-Liquid Feed) Non-interfering Bed Limiter Gas Injection Support Plate ‘Vapor Distributor” Vapor In e Notes: For Height Dimensions Associated With Each Device, See Table 5aa/NdD “oVeH PINbIT, oe oz oo ° —_—ToooOTO oo TOF Xxnyou [e001 VISd 00€ ’oU/*9! s6uly ited .Z oz ma NN, 3 soInqUIsiG Ued dug MON $ ax 2m aa ex a3 aS ss 2 L AONAIOISSa NO LOWd ~”? ; 7g NVOMINGIS. 3AVH NvosoinarisidFigure 2 TYPES OF TRAYS (SCHEMATIC) Sieve a) Fr vatve 7 —— (ee) ho} ae Fully = eo (1-0) Co) Standard Vacuum ap cap EXXON RESEARCH AND ENGINEERING COMPANY — FLORHAM PARK, Nu.EXON XX @reaineernc DESIGN PRACTICES FRACTIONATING TOWERS Tecton EXXON PACKING AND GRID ms mace PROPRIETARY INFORMATION — For Authorized Company Use Onty Pate Nov, 1989 Figure § TYPICAL LIQUID DISTRIBUTORSb-v pajeeg-eiqnog yosyi5 ev YOSHI, Ledhy yooy, Hae JOUNN A USSHID v edk| yooy Sap shell aajeA e e eADDENDUM 2.4 SECTION 2 DISTILLATION SELECTION AND DESIGN OF COMMERCIAL FRACTIONATION EQUIPMENT BY P. W. BECKER AND R. K. NEELD EXXON RESEARCH AND ENGINEERING COMPANY Although Chemical Engineering technology has grown tremendously in the past ten years, unit operations stil] remain at the heart of the chemical and petroleum industries. And, within the unit operations area, distillation remains the chief means of the various physical separation techniques. Distillation as an art is old; but distillation as a science is much newer, and a proportionate share of the R&D effort involved in developing new technology has been and will continue to be expended in this area. The purpose of this handout is to discuss what some of this technology is, especially from the standpoint of the design engineer. First, why is it important. Second, what changes have taken place in recent years. Third, what tools do the designers have to work with and how do they go about selecting and designing conmercial distillation equipment. Importance of Distillation The large annual investment in distillation equipment is a major reason why work in the fractionation area is important to us, and why we undertake research and development work; spending both time and money trying to perfect our technology in this area. For example, during the past several years, Exxon Corporation alone has spent over ten million dollars annually for towers and trays. Obviously, then, the possibility of even small improvements in the selection, design, and performance of our distillation units can be readily justified. A second reason for a continuing effort in this area is that distillation is usually the final processing step that determines whether on-spec products are being produced. And, thirdly, the purity specifications for both new and existing products are constantly being raised, thereby requiring more accurate tower design techniques. Furthermore, we could cite a variety of system physical properties, yield-purity relationships, liquid and vapor rates, temperature sensitive materials, and fouling requirements that create special problems affecting hardware design. Since the oi1 embargo of 1974, and the subsequent quadrupling of oi] prices, another important consideration has arisen--that of conserving energy. Distillation is a large energy consumer in refineries and chemical plants. Heat must be put in at a high temperature level at the bottom, and removed at a lower level at the top. This introduces thermal inefficiency, and necessitates the use of lower reflux and lower pressure drop internals. 08159001.c¥SDESIGN PRACTICES EXON | FRACTIONATING TOWERS Section Page EXXON DEVICE SELECTION AND BASIC CONCEPTS WA 29 0147 Date ENGINEERING PROPRIETARY INFORMATION — For Authorized Company Use Only Dec. 1968 e aes ‘TYPES AVAILABLE (PACKING, GRIDS, BAFFLE SECTIONS, DUALFLOW TRAYS) (Cont) Figure 20 PHOTOS OF DUMPED PACKINGS PALL RING NORTON METAL INTALOX (MTP) a NUTTER RING ‘There are a number of other packing types available but they are not widely used in Exxon towers and therefore ‘not been included here. 4) b) Structured packings (also called ordered packings). These devices are fabricated in bundles from crimped sheet ‘metal and installed in the tower in layers having a fixed orientation. Since they provide more surface area per unit e volume they are more efficient than random packings. However, they cSBTZ-4 Trn@s as much, Since the crimp height, can be changed, the capacity, efficiency, pressure drop and cosl car also be varied Thus, the optimum choice must be determi ‘economies of the contacting doves ayalabe, ney provide the lowes pressure cro per theoretical stage of contacting as | well as the best capacity/etticiency combination. This feature makes them especialy attractive in vacuum towers ‘There are several suppliers including: Flexipac by Koch Engineering, Gempak by Gitsch, Intalox Structured by Norton, Montz by Nutter Engineering, and Meliapak by Sulzer. One of these devices, by Koch Engineering, is shown in Figure 21. Figure 21 PHOTOS OF STRUCTURED PACKING {BY KOCH ENGINEERING) ISOMETRIC VIEW EXXON RESEARCH AND ENGINEERING COMPANY — FLORMAM PARK, NJ.Why New Devices Outperform ‘The Bubble Cap Tray We must also remember that the variety of equipment (or hardware) available to handle these process-oriented problems is also constantly growing and being steadily improved. For instance, in the early 50’s and for many years before, the bubble cap tray was the workhorse of the distillation field. This is no longer true, and we now have: * devices with greater capacity and lower pressure drop than the bubble cap, * devices with efficiencies that are equal to or better than the bubble cap, * devices that are lower cost, * devices that possess almost equal flexibility to handle varying liquid and vapor loads without loss in effectiveness, and finally, © devices that are easier to maintain. Tray Performance Diagram And Design Limitations A number of physical boundary conditions limit our final tray design. The final tray design should eliminate or, at the very least, minimize the impact of each of these limitations. The correlations available for calculation a "performance diagram" for each contacting device vary from company to company. Contractors have correlations obtained from various sources, tray vendors generally have developed their own, as some companies have also, while other companies may use Fractionation Research Inc. or the literature sources. Regardless of the source, these design limitations are calculable, and the various correlations themselves will not be discussed. The physical boundary conditions include: Weeping The weep point is that vapor rate at which liquid starts to leak through the tray openings. It is not necessarily the lower operation limit for good tray efficiency. For systems with high liquid-to-vapor ratios, a smal] amount of liquid bypassing will not seriously reduce tray efficiency, providing it is only a small fraction of the total liquid on the tray. Dump ing Dumping is excessive leakage of liquid through the tray openings. It is characterized by a sudden and significant drop in tray efficiency. The minimum vapor rate for acceptable tray performance is equal to or greater than that at which dumping occurs. 08159001.c¥5Blowing Blowing is a fine dispersion of fog of tray liquid which becomes entrained to the tray above. It is caused by excessive vapor velocity through the tray openings at relatively low liquid loadings. The high hole velocity is caused by a high vapor loading or low percent hole area, or both. Blowing results in poor contacting since it lifts the liquid or froth phase off the tray and the vapor phase thus becomes the continuous phase. This limitation does not necessarily result in flooding. Flooding is an unstable condition in which the liquid height on the tray and in the downcomer builds up until the tower is essentially full of dense foam. The two principle causes of flooding are, first, excessive entrainment and, second, excessive downcomer filling. Entrainment Flooding or Jet Flooding Entrainment or "jet" flood is a condition where a vapor handling limitation is caused by the carryover or "jetting" of liquid droplets from one tray to the tray above because of excessive vapor velocity through the tower free area. This represents the maximum capacity of the tray at a given tray spacing. Excessive Downcomer Filling High tray pressure drop or insufficient disengaging of vapor in the downcomer results in a buildup of froth in the downcomer and eventual tray flooding. It can occur at any liquid at any liquid rate if inadequate domcougr clearance, inadequate downcomer area, or inadequate tray spacing is provided. iguid Gradient Liquid gradient may also cause problems. The change is depth of liquid on a tray, from the inlet toward the outlet, is the liquid gradient. Depending on the type of tray (that is, the resistance to flow) ad the motion of the vapor, the static head represented by the liquid gradient may furnish anywhere from a negligible to a major part of the driving force that moves the liquid across the tray. Gradient problems are worst on bubble cap trays and least on jet trays. loci sen The liquid velocity into the downcomer must be low enough to allow vapor to disengage and travel up and out of the downcomer against the flow of incoming froth. If this velocity is exceeded, the increased downcomer level 08159001..crsdue to excessive aeration may cause flooding of the tray. In addition, the vapor in the froth mixture may be of such magnitude that when it disengages on the tray and result in premature jet flooding. timate Capacity Ultimate capacity is another limitation that is occasionally encountered. The ultimate capacity is the upper limit of vapor loading which the tower can handle. It depends mainly on the physical properties of the system. This vapor load cannot be increased by changing the tray design or increasing the tray flooding occurs first; but, in certain cases, an ultimate capacity limitation may be reached before jet flooding. Thus, each tray design should always be checked for ultimate capacity. Flexib) yr Turndown Ratio Flexibility is the ratio of maximum to minimum vapor loads which bound the range of operation conditions over which the tray will perform satisfactorily; i.e., tray efficiency remains above, roughly, 90 percent of its maximum value. The flexibility of a typical tray is depicted in Figure 1. We see that tray efficiency remains relatively constant in the range of loadings from 40 to 90 percent of the flood point. Below 40 percent of flood, efficiency decreases due to weeping, dumping, and poor contacting due to low interfacial area. Above 90 percent of flood, efficiency rapidly. drops off due to high entrainment. If better turndown that this is required, a special device might be required. Therefore, it is important that you, as a designer, carefully determine the minimum loading conditions that the tower will actually see. Quite often, arbitrary or unrealistic turndowns are specified, and the cost of the tower increases appreciably; for example, shifting to a costlier device. Now that we briefly reviewed the various tray limitations, let’s look at the various types of trays available to us. The contacting devices can be broken into three broad categories. These include: downcomer Type Trays, Downcomerless Type Trays, and Packing. Category 1 - The Downcomer Type Trays which include bubble cap, sieve, and valve trays. Some of the major characteristics of each device include: tegory ncomer_Type Trays + Bubble caps - There are many, many types of bubble cap tray designs available to the industry. And, surprisingly enough, there is probably very little difference in the performance of those designs that might be called the best. The better designs are characterized by being relatively high in efficiency, in capacity, and in flexibility. In general, however, they do possess a higher pressure drop than the other devices that we'll talk about. But, their biggest deficiency is that of cost. Other types of trays are as much as 50 percent cheaper than bubble cap trays; and this is, perhaps, the main reason why bubble caps have lost their popularity. 08159001..c¥5Sieve trays - We find that they are currently the most widely-used of all the trays available. A typical, modern sieve tray is shown on the next Slide 6. This tray exhibits good capacity, excellent efficiency, and good flexibility--that is, it will operate quite efficiently at loadings which are one-half to one-third of design values. And, perhaps most importantly, it is the cheapest device currently available. One reason for the increasing popularity of sieve trays can be traced to the work of an organization called Fractionation Research Inc. (FRI). This organization has been in operation for over 20 years as a non-profit, industry-sponsored research cooperative. It has published performance data on almost all of the important contacting devices currently available. These data have allow not only pretty good comparative evaluations to be made, but have also allowed design procedures to be established. Valve trays - These trays contain proprietary devices manufactured by three concerns: Koch Engineering, F. W. Glitsch & Sons, and Nutter Engineering. These trays consist of a number of discs (or valves) suspended above the holes of a perforated tray at a height which depends upon the vapor flow rate to the hole. The trays themselves are characterized by a high turndown ratio. Capacity and efficiency are both high, and are similar to sieve trays, while their cost averages about ten percent more than sieve trays. They are generally used where wider flexibility than that obtained from sieve trays is required. By the way, this is a "two-pass" valve tray. That is, the liquid flows from left to center, or from right to center. On the trays directly above or below this one, liquid flows from the center to either side. Jet trays are similar in construction to a standard sieve tray except the punched holes are replaced by inclined tabs. The tabs are punched from the tray deck and inclined to the plate at a typical angle of 20° from the horizontal. This tray is characterized by having a high liquid and vapor handling capacity. Liquid handling capacity is high because the vapor assists or “pumps” the liquid across the tray. The tray is used in heavily-liquid loaded services where multi-pass convential trays are normally required. For most distillation towers (where liquid rate is normally low), the tray is not generally used since its capacity advantage becomes marginal, and its efficiency is less than a sieve or valve tray. V-Grid tray - A typical V-Grid tray is a convential crossflow tray. Distributed throughout the tray deck are V-Grid units, which can best be described as rectangular valves which are always in the fully-open position. The main difference between the operation of the V-Grid tray and a typical sieve tray is in its vapor flow patterns. While vapor always flows in a vertical direction through the sieve tray, it must flow horizontally through the V-Grid units. Because V-Grid units are laid out on a rectangular (rather than triangular) pitch, vapor jets leaving adjacent units oppose each other, thereby promoting increased turbulence and mass transfer. 8189001 .crse The operation of the V-Grid was compared to that of a sieve tray during a laboratory test, program conducted by Nutter at their facilities in Tulsa, Oklahoma(1). These data indicated that this increased action on the V-Grid tray resulted in its having somewhat higher fractionation efficiency than a sieve tray. Also, liquid did not weep as readily through the V-Grid tray, which indicated that V-Grid has a better turndown ratio than the sieve tray although certainly not as good as a capacity advantage over a sieve tray under most conditions, it did exhibit higher entrainment than the sieve tray at very low liquid rates. Thus, at moderate and high liquid rates, the V-Grid tray’s performance is at least equivalent to, and possibly somewhat better than, that of a sieve tray. At low liquid rates, however, V-Grid trays should be used cautiously since they may entrain excessively. The V-Grid tray should also be particularly attractive in fouling services. The violent tray action (caused by the opposing vapor jets) should make the V-Grid tray a self-cleaning device. Nutter claims that their trays have been used successfully in services which were previously plagued with tray fouling problems. ategory ‘The Downcomerless Trays © Dual flow is a generic term and describes, basically, a downcomerless sieve tray in which both liquid and vapor flow takes place through the same holes. This could well be the sieve tray you just looked at with e the downcomers completely blocked off and the area Filled with more of the perforated sheet that comprises the active area of the sieve tray. © Ripple tray - This is a proprietary device that’s been patented by the Stone and Webster Engineering Company. They fabricated this device by crimping a standard flat dualflow tray into sine waves of a desired frequency and amplitude. We have some data on the performance of both the dualflow and ripple tray and have found they possess very high capacity and moderate efficiency. However, the flexibility or turndown ratio over which good efficiency can be obtained is generally poor. As a matter of fact, unless the tray is properly designed for the load at which it is to operate, it won’t do the job. It is normally considered only when higher capacity is desired and Some sacrifice in efficiency and turndown ratio can be tolerated. This ‘is the case in some debottlenecking studies. (1) The results of this program were discussed in detail in A.I.Ch.E. Preprint 49c (Ammonia Stripping Efficiency Studies), Sixty-Eighth National Meeting, March-April, 1971 by D. E. Nutter. 8159001. cysCategor: in Four of the more familiar packings that are widely used today are the Intalox Saddle, the Pall ring, the Raschig Ring, and the Berl Saddle. Both the Raschig Ring and the Beri Saddle, while still used in some limited services, have been largely replaced by the Pall ring and the Intalox Saddle. This is due to the higher capacity and efficiency of these newer packings. The Raschig Ring is nothing more than a short cylinder. When made from ceramic material, its walls are usually quite thick. The Pall ring, on the other hand, has thin, slotted sides with fingers that protrude inward. This permits much more of the ring’s surface to be used in the mass transfer process by producing more turbulent contacting and reducing stagnant zones. Likewise, the new Intalox Saddle is more efficient than the older Berl Saddle it has largely replaced. On an overall basis, i.e., efficiency, capacity, turndown, and cost, we have found the two-inch metal Pall ring to be the most economic’packing to use in most of our distillation services. Packed towers are primarily used in the following applications: vacuum systems where low pressure drop is critical, + debottlenecking existing trayed towers, * light ends towers at high liquid loadings, * absorption and stripping services (high liquid loadings). In addition, packing is also used frequently in small diameter towers (under ~ three feet) and in corrosive services where a ceramic material of construction is required. Are These All of the Contacting Devices Available The trays and packings discussed previously will usually fill more than 95 percent of the designer’s needs. However, there are certain applications where some of the more non-conventional or high cost devices are justified. Although there are many, only a few more commonly-used ones are described here. These are: + The MD Tray - This tray is a perforated fractionating device which features a multiplicity of narrow downconers distributed over the entire tray surface. Hence, the name "multiple downcomer". The Linde(2) tray is a high capacity device which makes it especially attractive for debottlenecking a tower. While its efficiency is somewhat lower than a (2) Detnicki, W.V., and Wagner, J. L., “Performance Characteristics of Linde Multiple Downcomer Distillation Trays", Preprint 24c A.I.Ch.E. Sixty-Second Annual Meeting, November 16-20, 1969. osisso01 cysstandard cross-flow sieve tray, its high capacity may mean that extra trays can be added to a given tower or the reflux rate increased somewhat to offset any fractionation debit. For a specific case, you should contact Linde to determine whether the MD tray is attractive for your application. Sulzer Packing - This packing is manufactured by Koch Industries, Inc. The packing is a woven wire fabric material which is completely self-wetting. The packing is a high efficiency device (i.e., low HETP) with a very low pressure drop. It has special application in distill ing heat-sensitive materials under vacuum conditions. Its main disadvantage for more gengral use is its cost, being about $450/ft3 (as compared to about $20/ft3 for stainless steel two-inch pall rings). Pe Kontakt Tray - This device is made from pieces of expanded metal and contains a number of intermediate weirs. In addition, the metal is expanded in such a manner that the vapor helps "push" the liquid across the tray. Because of this feature, the tray is claimed to have a high capacity and is, therefore, recommended for debottlenecking as well as new designs. Unfortunately, due to tis newness in the U.S., relatively little data is publically available for comparison purposes. The vendor should, however, be contacted for more specific details and recommendations on the application of this tray(3). What is Involved in Tower Design? Having discussed some of the characteristics of the equipment that designers, might use, let’s briefly go through the steps involved in designing a distillation tower. With the advent of computers, this has becone more rigorous in recent years. The steps involved are summarized as follows: 1, Define the key separations - Product specifications or degree of separation can be given to us in a number of ways; perhaps as a flash point or an amount of material boiling below or above a certain temperature. On simpler towers, where individual components can be identified, the amount of purity is given, We must, by some means, tie this into a yield and quality description. Although we don’t need, at the start, to have a complete definition of the products (that is, we don’t need a complete material balance on all components), we do need to know the key requirements of the separation, that is, the yield and the purity level of our products. For further information, contact Ferrostaal Overseas Corporation, 17 (3) Battery Place, New York, New York. (212) 422-6535, Attention Mr. Alfred B. Walter, Manager. 08159001..c¥s,2. jate vapor-liqui ilibri enthalpy data - A number of basic data correlations are available to use. At last count, Exxon had over a dozen different correlations or techniques in use for different types of problems. One of our big jobs as designers is to define the limits and proper application of each of these correlations. 3. Calculate heoretical plates requi! iff f1 ios - The calcuTation techniques for determining the right combination of theoretical trays and reflux ratio has improved tremendously in recent years, and these techniques are limited now only by the accuracy of the basic data. Today, we have plate-to-plate programs that can handle up to 150 components, five sidestreams, and five heat removal sections or pumparounds. A complete printout gives us tray and produce vapor and liquid rates, temperatures, compositions, and certain physical property information.” If we run several cases with different numbers of trays and different reflux ratios, we can then define the shape of the theoretical trays vs. reflux ratio curve. Before we can select the optimum combination, we need to: 4. Estimate the tray efficiency - Lacking specific data on a similar system, there are a number of correlations available to us to do this, ‘the most fundamental being the A.I.Ch.E. method for bubble cap trays. Since this correlation was developed from bubble cap data on binary systems, most of them aqueous solutions, there is some question as to its applicability to the multicomponent hydrocarbon systems frequently encountered in the petroleum industry. Exxon has, therefore, derived our own correlations for predicting sieve tray efficiency. With our estimated plate efficiency, we can now go back and calculate the actual number of trays that corresponds to each reflux ratio. We now have the basis for an economic balance to determine which is the proper combination to use. In general, we have found that reflux ratios in the range of 1.2 to 1.5 times the minimum are usually optimum. 5. Define the minimum and maximum feed rates that our tower must handle - Tt is important that the designer be given the range of loadings over which his tower must operate, and this is a number that is frequently not easily obtained. In general, an assumption of a2 to 1 rangeability fits most normal operations. When greater flexibility than this is required, the degree of flexibility becomes one of the key items in determining both the type of tray to use and how it should be designed. The next step in designing towers should be: 6. Selecting the best tray for the given set of process conditions. Choosing the best tray type goes back to our earlier discussion on the merits of the various trays. For about 90 percent of new designs, sieve trays have proved adequate to cover both the range of loadings and the various services required. Of course, such items as coking tendency, allowable pressure drop, flexibility requirements, and, in the case of bottleneck removal, required capacity and/or efficiency will have sone bearing on tray selection. By and large, however, sieve trays have og1sg001.cys-10- become the design engineer’s first choice among the available contacting devices. If packing is used, it is hard to beat the two-inch Pall ring. 7. Tower sizing and tray hydraulic calculations complete our tower design. Most companies have their own design methods for this phase of tower design; and, in almost all cases, towers are designed to flood by entrainment; that is, liquid is carried from one tray to the tray above by the ascending vapor. But, it is also essential to check the overall tray pressure drop, downcomer liquid velocity, and downcomer Filling to assure that the trays will be hydraulically stable. While the hydraulic calculations are pretty well standardized, tray capacity correlations are, by and large, of the generalized empirical variety. Principal variables governing the allowable vapor velocity are: liquid and vapor density, liquid surface tension, viscosity and rate, and certain tray hardware factors; namely, tray spacing and percent hole area. 8. Process = In general, tower control schemes were pretty well standardized until a few years’ ago when so-called super fractionators, and by this we mean towers with large nunbers of trays and high reflux ratios, appeared on the scene. This has given rise to a number of studies of control schemes involving lag time estimates, temperature-composition sensitivity, (which, incidentally, is usually poor), and chromatographic analyzer applications. In addition, small analog computers seem to offer promise in this area to optimize utilities and to standardize product output of these towers since large lag times make manual control difficult. Tower control issues may thus influence the final tower design. 08159001.cysSECTIONS 3 & 4 HEAT TRANSFER AND HEAT EXCHANGERS What We'll Cover: SECTION 3 © Heat Transfer Theory-Review * Relation of Heat Transfer Theory to Shell and Tube Heat Exchangers * Design of a S&T Exchanger—Procedure Outline SECTION 4 * Design Features and Parameters of Shell and Tube Exchangers BASIC HEAT TRANSFER CONCEPTS * Flow of Heat Behaves Like Flow of Fluids and Flow of Electrons |, Driving Force Rate a K « SAVING FORE (General) . Pressure Drop Resistance 10+ penta (Electricity) Qak (Fluids) __ Temperature Difference oo Resistance Ae)COMPARISON WITH FLUIDS 2 Fluids: (2)" = «« Pe-Pr (Section 6 wit! A ( & 7 Treat Fluid Flow) D Dat. Toth Heat: 2 = 1-6 TE FiDs HEAT @ = Volume/Second Q = Btu/Hour P,, Py = Higher, Lower Ta, Ty = Higher, Lower Pressures Temperatures A = Area Available for Flow Ry = Total Specific Resistance fe A = Area Available for Flow (4 )= Number of Fluid of Heat Flow Resistance Units BASIC HEAT TRANSFER EQUATION Qay-.y.ar A Ry Ry Ry = Total Resistance, HeFT?-°F/Btu ie = Total Conductivity = U, Btu/Hr-Ft?-°F Det-uar Q = U,+ A> AT Btu/Hr U, Is Referred to as the “Overall Heat Transfer Coefficient”TOTAL RESISTANCE TO HEAT FLOW—HEAT EXCHANGERS * There Are Two Areas Through Which Heat Must Flow: the Inside Tube Area and the Outside Tube Area. Resistance Occurs at Both Areas. * The Industry Standard Reference Area Is the Outside Tube Area, INDIVIDUAL COMPONENTS OF THE TOTAL RESISTANCE Inside Film Resistance = Ri. = Fi( %) Inside Fouling Resistance = 14, = 5, ( *) Tube Wall Resistance = 1, = é Outside Fouling Resistance = r, Outside Film Resistance = R, Ro + fo + fw + fo + Ry = Rr = 7 &, = Wall Thickness, Feet K, = Thermal Conductivity, Btu/Hr-Ft?- zt f= Resistances, Hr-Ft?-°F/BtuTYPICAL RESISTANCE VALUES Very Low _Typical_ Very High Film Resistances (Each) 0.00050 0.004 0.04 (inverse = h) (2000) (250) (25) Fouling Resistance (Each) 0.001 0.002 0.01 Inverse (1000) (600) (100) Wall Resistance 0.000030 0.00027 «(0.00049 Inverse (32,000) (3760) (2030) Total Resistance 0.00303 0.01227, (0.10050 Inverse (330) (1) (10) THE CONTROLLING COEFFICIENT * Frequently one of the Two Film Coefficients determines the Value of the Overall Coefficient: Outside Coefficient, h, = 75 75 150 Inside Coefficient, hi. 1000 3000 1000 R, = 0.01333 0.01333 0.00667 Rig = 0.00100 0.00033 0.00100 Twit le +f. = 0.00070 0.00070 0.00070 Ry = 0.01503 0.01436 0.00837 U= 665 69.68 1195 Improvement = Base +46% = +80% * Hence h, is the ‘Controlling Coefficient”, and Efforts to Improve Exchanger Performance Should Concentrate on This Side of the Exchanger.A. TEMPERATURE DROPS ACROSS THE RESISTANCES + Temperature Drop Across Each of the Resistances is Directly Proportional to Each Resistance. * For Example, If T, = 200 and T, = 80, Then Total Temperature Drop = 120°F, and: Temperature Drop 0.01333 R, = 0.01333 776 = 9.01333 Rig = 0.00500 29.1 aad fw = 0.00030 17 To + fo = 0.00200 1 Ry = 0.02063" 120°F B. TEMPERATURE DROPS ACROSS THE RESISTANCES A Useful Concept is Heat Fiux = (2) = Bt Q=U+A+ (12-1) =U A+ aT Then aT = = (2) = Fux. Resistance Q AT a U-A 120 In Previous Slide, If A = 4000 Ft? Then p= Ry = 002063 = 5817 —BW. and AT Across Ry = 5817-0.01333 = 75°F ? as Shown on That Slide,BACK TO BASICS + We've Looked at Basic Theory, and Discussed Q = Up + A + AT. In Refinery Work We Usually Know Elther Q or A, and Need to. Calculate the Other Value. How Do We Do It? + Either Question Requires Calculating Ug and AT. + We'll Talk About U, Later, First Let's Discuss AT, The ‘Temperature Driving Force. + Note that Capital Letter T Denotes the Hot Stream, While Lower Case t Denotes the Cold Stream: Ty = Hotin —T,= Hot Out ty) =Coldin ty = Cold Out 310 FLOW PATTERNS AND TEMPERATURE DRIVING FORCE A. Cocurrent (Parallel) FlowFLOW PATTERNS AND TEMPERATURE DRIVING FORCE 8. Countercurrent Flow i Double Pipe Exch. 4 h-hh It, Cross When bot FLOW PATTERNS AND TEMPERATURE DRIVING FORCE Mixed FlowFLOW PATTERNS AND TEMPERATURE DRIVING FORCE ©. Transverse Flow : y”" we" Tete" a, este ay 7 TEMPERATURE DRIVING FORCE * From the Preceeding Slides, It is Clear That Some Sort of ‘Average Driving Force Must Be Used in Design Calculations. What is This Average? * The Average Is Called ““The Effective Mean Temperature Difference”, or MTD,. * For True Countercurrent and True Cocurrent Flow, the Effective Driving Force Equals the Log Mean Average of the Two Extreme (Largest and Smallest) Deltas, ATT, aT, = LMTD = ATS ATs This is Precisely True Only When the Heat Release Curves are Straight Lines. Otherwise it is an Approximation.TEMPERATURE DRIVING FORCE What About Mixed Flow: Shell and Tube Exchangers? ‘The Complex Flow in These Units was Analyzed Mathematically Many Years Ago, Resulting in Rigorous Equations for a Correction Factor, F,. This is Multiplied by the LMTD to Give the Correct MTD,. MTD, = F, * LMTD Equations are Valid Only When Heat Release Curves are Lin Similar Relations are Available for Transverse Flow (Air Fin Coolers, For Example). a6 CALCULATION OF F, * Depends on the Number of Shells in Series (“Shell Passes”) * The More Shells One Has in Series, the Closer F, Approaches 1.0 * Typically the Minimum Acceptable Value of is 08 + What Exactly Do We Mean by “Shells in Series” or “‘Sheli Passes"? a7CALCULATION OF F,—SHELL PASSES ONE PASS % 4 : th 7 Ta te Max 257, CALCULATION OF F,—SHELL PASSES ‘THREE PASSES Max t,>>T, a0CALCULATION OF F,, * Complex Equations Simplified to Charts * See TEMA Section 7, Pages 112-127 or Exxon DP IX-D, ppS3-pp58 * Applicable Only to Linear Heat Curves Min. for Design 1.0 “Temperature Efficiency” 320 CALCULATION OF F,, F, (1 Shell) = <0.8 (Unreadable on Chart)-Unacceptable F, (2 Shells) = 0.95 Use Two Shells, oatCALCULATION OF F,, * Since This Technique is Applicable Only to the Case of Straight-Line Heat Release, How Do We Estimate Number of Shells and MTD, for Other Cases? NON-LINEAR HEAT RELEASE—MTD, ‘SUGGESTION FOR COMPLEX CASES. SUCH AS REFORMER FEED/EFFLUENT —Plot Tvs. Enthalpy — Step Off to Get Minimum Number of Shelts — Calculate MTD, for Each Shell (Discuss Later) ‘Temperature Need Minimum ot 5 ShellsNON-LINEAR HEAT RELEASE—MTD, SUGGESTION FOR CONDENSERS —Plot the Condensing Curve Assume Cold Side is Line: Draw in Cold Side Flow Pattern It Two Shells, Assume Equal Duties ‘TOP SHELL BOTTOM SHELL Oesuperheatinge}-—————condenaing +} Sub-- a MTD, FOR CONDENSERS (Cont’d.) * Calculate the LMTD for Each Zone, Assuming That the Cold ‘Temperature in Each Zone Is the Average of the Inlet/Outlet Cold Temperatures of the Shell in Which the Zone Occurs (See Graph) * Then Weight the Overall MTD, as Follows: Qtotal MTD, (Weighted) = Gone1 _Gzonea — Guanes |, Gzoned tmTo, * LMTo, * LMTD, * LMTD, a2HEAT TRANSFER COEFFICIENTS + Film Coefficients are Relatively Easy to Estimate: They are a Function of + Reynolds Number _ C, + Prandtl Number Shi ‘See Addendum 4.01 + Similarly, Pressure Drop is a Function of Reynolds Number and Length of Flow Path. ? 26 HEAT TRANSFER COEFFICIENTS (Cont’d.) * The Handouts Just Examined are Suitable ONLY for Estimates of Coefficients. ‘* For Detailed Coefficients on Which to Base the Purchase of an Exchanger, Detailed Computer Calculations are Necessary. * Detailed Computer Calculations Examine the Effects of Many Other Parameters, Particularly Shell-Side Effects Such as Channelling and Baffle Leakage. az7EXCHANGER DESIGN PROCEDURE First Need to Know: * Permissible Tube Sizes—Diameter, Gauge, Length. (Frequently Set by Refinery Maintenance Department) The Appropriate Tube Material for the Service The Allowable System Pressure Drops for Each Stream. EXCHANGER DESIGN PROCEDURE (Cont’d.) (1) Assume an Overall Heat Transfer Coefficient (U,) and Calculate Area, (2) Using the Required Tube Size and Length, Calculate the Number of Tubes, (3) Using a “Reasonable” Tube-Side Velocity (2-15 Ft/s), Calculate the Tube Side Flow Area Required: _ fs . (4) Determine the EVEN Number of Tube Passes Which Will Most Closely Approximate the Needed Flow Area 329